Dissertations / Theses on the topic 'Artificial Intelligence and Signal and Image Processing'
To see the other types of publications on this topic, follow the link: Artificial Intelligence and Signal and Image Processing.
Create a spot-on reference in APA, MLA, Chicago, Harvard, and other styles
Consult the top 50 dissertations / theses for your research on the topic 'Artificial Intelligence and Signal and Image Processing.'
Next to every source in the list of references, there is an 'Add to bibliography' button. Press on it, and we will generate automatically the bibliographic reference to the chosen work in the citation style you need: APA, MLA, Harvard, Chicago, Vancouver, etc.
You can also download the full text of the academic publication as pdf and read online its abstract whenever available in the metadata.
Browse dissertations / theses on a wide variety of disciplines and organise your bibliography correctly.
1
Lambert, T. "Digital Enhancement Techniques for Underwater Video Image Sequences." Thesis, Honours thesis, University of Tasmania, 2005. https://eprints.utas.edu.au/253/1/tristanlthesis.pdf.
Full textAbstract:
Due to concern about the current state of the world's oceans, several large scale scientific projects have begun to investigate the condition of our oceans. These projects are making use of underwater video sequences to monitor marine species. The move to using underwater video monitoring introduces labour intensive manual processing techniques. This leads to the need for an automated system capable of processing the data at a much greater speed. This project investigated whether the development of suitable image processing techniques could be used for pre-processing underwater images from a fish farm and locating fish within these images using computer vision techniques.
Using underwater images leads to some serious problems when compared to images from a clearer environment. Visibility in an underwater environment is poor, even when using state of the art equipment. After reviewing the broad field of computer vision and current underwater projects, an image pre-processing system was developed in MATLAB using suitable image processing and analysis techniques.
The application developed was able to successfully locate an acceptable number of fish within the underwater images. The project demonstrated that automated analysis of underwater video images is needed and is possible. Automatic processing of large quantities of video image sequences will be of great benefit in the future. It will allow scientific researchers to study the ocean environment and its species more effectively. Pre-processing is an essential component of the overall process that will lead to automation of underwater video data analysis for marine science applications.
2
Fearn, RC. "The Generalisation Ability of Neural Networks." Thesis, Honours thesis, University of Tasmania, 2004. https://eprints.utas.edu.au/119/1/thesis.pdf.
Full textAbstract:
Neural Networks (NN) can be trained to perform tasks such as image and handwriting recognition, credit card application approval and the prediction of stock market trends.
During the learning process, the outputs of a supervised NN come to approximate the target values given the inputs in the training set. This ability may be good in itself, but often the more important purpose for a NN is to generalise i.e. to have the outputs of the NN approximate target values given inputs that are not in the training set.
This project examines the impact a selection of key features has on the generalisation ability of NNs. This is achieved through a critical analysis of the following aspects; inputs to the network, selection of training data, size of training data, prior knowledge and the smoothness of the function.
Techniques devised to measure the effects these factors have on generalisation are implemented. The results of testing are discussed in detail and are used to form the basis of further work, directed at continuing to refine the processes involved during the training and testing of NNs.
3
Zehmeister, MS. "Development of an Illumination Identification System for the AIBO Robot." Thesis, Honours thesis, University of Tasmania, 2005. https://eprints.utas.edu.au/247/1/mszThesisFinal.pdf.
Full textAbstract:
The Four Legged League is a division of the RoboCup initiative that uses Sony AIBO robots to further robotics research. Most participants implement vision systems that use the colour of objects to perform identification. Calibration of the colour classification system must be done manually and any changes to the lighting of the environment after calibration reduces the accuracy of the system, often to a point at which the robot is effectively blind.
This study investigates the relationships in the colour data of image pixels between different lighting conditions in an effort to identify trends that can be used as the basis of a rule-based system. The aim of the system is to identify the current lighting level as one of a set of known conditions.
The proposed system uses the colour data of image pixels and
information about the AIBO's location and orientation to identify lighting levels, allowing a vision system to switch to an appropriate pre-cofigured calibration.
4
Hall, DJ. "Flexible Robot Platform For Autonomous Research." Thesis, Honours thesis, University of Tasmania, 2005. https://eprints.utas.edu.au/249/1/djhThesis.pdf.
Full textAbstract:
The field of mobile robotics is receiving increasing levels of research. However, the simulation tools which are utilised in the creation of new mobile robot algorithms can produce algorithms which do not work in the real world. In order to try and minimise this problem a flexible robot platform has been created which allows the testing of a variety of algorithms. The platform facilitates the testing of algorithms normally only simulated by allowing algorithms to be easily tested in the real world. Utilising the flexible robot platform for testing algorithms allows higher quality research, as algorithms can be assessed with physical evidence.
5
D'Alton, S. "A Constructive Neural Network Incorporating Competitive Learning of Locally Tuned Hidden Neurons." Thesis, Honours thesis, University of Tasmania, 2005. https://eprints.utas.edu.au/243/1/D%27Alton05CompetitivelyTrainedRAN.pdf.
Full textAbstract:
Performance metrics are a driving force in many fields of work today. The field of constructive neural networks is no different. In this field, the popular measurement metrics (resultant network size, test set accuracy) are difficult to maximise, given their dependence on several varied factors, of which the mostimportant is the dataset to be applied.
This project set out with the intention to minimise the number of hidden units installed into a resource allocating network (RAN) (Platt 1991), whilst increasing the accuracy by means of application of competitive learning techniques. Three datasets were used for evaluation of the hypothesis, one being a time-series set, and the other two being more general regression sets.
Many trials were conducted during the period of this work, in order to be able to prove conclusively the discovered results. Each trial was different in only one respect from another in an effort to maximise the comparability of the results found. Four metrics were recorded for each trial- network size (per training epoch, and final), test and training set accuracy (again, per training epoch and final), and overall trial runtime.
The results indicate that the application of competitive learning algorithms to the RAN results in a considerable reduction in network size (and therefore the associated reduction in processing time) across the vast majority of the trials run. Inspection of the accuracy related metrics indicated that using this method offered no real difference to that of the originalimplementation of the RAN.
As such, the positive network-size results found are only half of the bigger picture, meaning there is scope for future work to be done to increase the test set accuracy.
6
Woolford, E. "Residual Reinforcement Learning using Neural Networks." Thesis, Honours thesis, University of Tasmania, 2005. https://eprints.utas.edu.au/250/1/ewFINAL_with_beginning_pages.pdf.
Full textAbstract:
A number of reinforcement learning algorithms have been developed that are guaranteed to converge to an optimal solution for look-up tables. However, it has also been shown that these algorithms become unstable when used directly with a function approximation system. A new class of algorithms developed by Baird (1995) were created to handle the problem that direct algorithms have with function approximation systems. This thesis focused on extending Baird's work further by comparing the performance of the residual algorithm against direct application of the Temporal Difference learning algorithm. Four benchmark experiments were used to test each algorithm with various values of lambda and alpha over a period of twenty trials. Overall it was shown that the residual algorithm outperformed direct application of the TD learning algorithm on all four experiments.
7
Kamenetsky, D. "A Comparison of Neural Network Architectures in Reinforcement Learning in the Game of Othello." Thesis, Honours thesis, University of Tasmania, 2005. https://eprints.utas.edu.au/252/1/dkThesis_Final_4.pdf.
Full textAbstract:
Over the past two decades, Reinforcement Learning has emerged as a promising Machine Learning technique that is capable of solving complex dynamic problems. The benefit of this technique lies in the fact that the agent learns from its experience rather than being told directly. For problems with large state-spaces, Reinforcement Learning algorithms are combined with function approximation techniques, such as neural networks. The architecture of the neural networks plays a significant role in the agent's learning. Past research has demonstrated that networks with a constructive architecture outperform those with a fixed architecture on some benchmark problems.
This study compares the performance of these two architectures in Othello - a complex deterministic board game. Three networks are used in the comparison: two with constructive architecture - Cascade and Resource Allocating Network, and one with fixed architecture - Multilayer Perceptron. Investigation is also made with respect to input representation, number of hidden nodes and other parameters used by the networks. Training is performed with both on-policy (Sarsa) and off-policy (Q-Learning) algorithms.
Results show that agents were able to learn the positional strategy (novice strategy in Othello) and could beat each of the three built-in opponents. Agents trained with Multilayer Perceptron perform better, but converge slower than those trained with Cascade.
8
Kelsall, A. "Flexible Shape Models for Marine Animal Detection in Underwater Images." Thesis, Honours thesis, University of Tasmania, 2005. https://eprints.utas.edu.au/248/1/afkThesis_FINAL.pdf.
Full textAbstract:
Many industries are benefiting from computer automation, however the area of image analysis is still limited. The process of finding a potential object in an image is hard in itself, let alone classifying it. Automating these tasks would significantly reduce the time it takes to complete them thus allowing much more data to be processed. This becomes a problem when data is collect faster than it can be analysed. Images and video sequences are captured for different purposes and need to be manually processed in order to discover their contents.
The fishing industry is a perfect example of this. A fish farm needs to know the average size of the fish in a ring. At present, this involves either manually taking a sample of fish from the ring and measuring them, or taking a series of stereoscopic images and manually tracing a sample of fish. By using active shape models, the process of tracing a fish sample can be automated. The Active Shape Model (ASM) Toolkit is an implementation of active appearance models, an advanced type of active shape model. The wrapper application that was written as part of this research allows a more streamlined process to input region data into the ASM Toolkit for searching. Once a sample has been matched, it is possible to use the key points around it to base further calculations on such as its size and weight.
The ASM Toolkit and the wrapper program demonstrate how the process of identifying a fish in an image can be automated and that it is possible to calculate the size and weight of fish. In an ideal manual test, the most effective model matched 68% of samples, and in the automated test matched 50% of the samples. If the program can run over several days collecting appropriate samples, the model will be able to match enough fish to estimate the average size and weight within a ring. It is shown that the types of samples used in training the model affects the performance more than the number of samples used.
9
Rochford, Matthew. "Visual Speech Recognition Using a 3D Convolutional Neural Network." DigitalCommons@CalPoly, 2019. https://digitalcommons.calpoly.edu/theses/2109.
Full textAbstract:
Main stream automatic speech recognition (ASR) makes use of audio data to identify spoken words, however visual speech recognition (VSR) has recently been of increased interest to researchers. VSR is used when audio data is corrupted or missing entirely and also to further enhance the accuracy of audio-based ASR systems. In this research, we present both a framework for building 3D feature cubes of lip data from videos and a 3D convolutional neural network (CNN) architecture for performing classification on a dataset of 100 spoken words, recorded in an uncontrolled envi- ronment. Our 3D-CNN architecture achieves a testing accuracy of 64%, comparable with recent works, but using an input data size that is up to 75% smaller. Overall, our research shows that 3D-CNNs can be successful in finding spatial-temporal features using unsupervised feature extraction and are a suitable choice for VSR-based systems.
10
Knights, MS. "Flexible shape models for image analysis in an automated lobster catch assessment system." Thesis, Honours thesis, University of Tasmania, 2007. https://eprints.utas.edu.au/3013/2/1_front_Knights.pdf.
Full textAbstract:
Management of fisheries is an evolving science combining multiple techniques and strategies. The involvement of the computer in industry management and research continues to grow. The area of image analysis is currently limited but continues to grow as computing equipment becomes faster and cheaper. Locating a particular object in an image and processing information about that object is a significant task that requires a great deal of processing power and finesse. The benefits of a functioning automated task that processes data on an object, such as a lobster, simply by processing an image of that object would greatly enhance the ability to manage a fishery with accurate, up to date data.
The Tasmanian Aquaculture and Fisheries Institute (TAFI) intend to create a lobster-sorting tray, which can be used on lobster fishing vessels as standard equipment. This tray would include functionality to take an image of the current lobster and estimate its sex and weight from pertinent measurements on the lobster. This research demonstrates that through the use of the Active Shape Modeller (ASM) these details can be identified and processed from an image of the lobster.
The ASM is used within an image analysis process, which can be fully automated, to draw out the required salient details of a lobster from an area of interest in the images. A series of experiments showed that the ASM was able to draw out and fully identify 77.3% images in a test set of 216 images. These images then had pertinent lengths and a sex estimated based on these measurements where 90% of the matched lobsters were sexed correctly.
11
Laungrungthip, Nuchjira. "Sky detection in images for solar exposure prediction." Diss., Lincoln University, 2008. http://hdl.handle.net/10182/855.
Full textAbstract:
This project describes a technique for segmenting regions of sky in an image from the remainder of the image. This segmentation technique is part of a method for predicting the solar exposure at a location of interest from a set of photographs. Given the latitude and longitude of the position and the direction and field of view of the camera it is possible to calculate the position of the sun in the image at a particular time on a particular day. If that position is in a sky region of the image then the location will be exposed to the sun at that time. Critical to the success of this method for determining solar exposure is the image processing used to separate the sky from the rest of the image. This work is concerned with finding a technique which can do this for images taken under different weather conditions. The general approach to separate the sky from the rest of the image is to use the Canny edge detector and the morphology closing algorithm to find the regions in the image. The brightness and area of each region are then used to determine which regions are sky. The FloodFill algorithm is applied to identify all pixels in each sky region. An extensive empirical study is used to find a set of threshold values for the Canny edge detector, applied to the blue colour channel, which allow successful identification of the sky regions in a wide range of images. Tests using different camera filters show that they do not usefully increase the contrast between the sky and the rest of the image, when a standard compact camera is used. The work reported in this thesis shows that this approach of finding edges to identify possible sky regions works successfully on a wide range of images although there will always be situations, such as when the image is taken directly into the sun, where manual adjustment to the identified regions may be required.
12
Punchihewa, Amal. "Synthetic test patterns and compression artefact distortion metrics for image codecs : a thesis presented in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Engineering at Massey University, Palmerston North, New Zealand." Massey University, 2009. http://hdl.handle.net/10179/1131.
Full textAbstract:
This thesis presents a framework of test methodology to assess spatial domain compression artefacts produced by image and intra-frame coded video codecs. Few researchers have studied this broad range of artefacts. A taxonomy of image and video compression artefacts is proposed. This is based on the point of origin of the artefact in the image communication model. This thesis presents objective evaluation of distortions known as artefacts due to image and intra-frame coded video compression made using synthetic test patterns. The American National Standard Institute document ANSI T1 801 qualitatively defines blockiness, blur and ringing artefacts. These definitions have been augmented with quantitative definitions in conjunction with test patterns proposed. A test and measurement environment is proposed in which the codec under test is exercised using a portfolio of test patterns. The test patterns are designed to highlight the artefact under study. Algorithms have been developed to detect and measure individual artefacts based on the characteristics of respective artefacts. Since the spatial contents of the original test patterns form known structural details, the artefact distortion metrics based on the characteristics of those artefacts are clean and swift to calculate. Distortion metrics are validated using a human vision system inspired modern image quality metric. Blockiness, blur and ringing artefacts are evaluated for representative codecs using proposed synthetic test patterns. Colour bleeding due to image and video compression is discussed with both qualitative and quantitative definitions for the colour bleeding artefacts introduced. The image reproduction performance of a few codecs was evaluated to ascertain the utility of proposed metrics and test patterns.
13
Nahodil, D. "Creating and Analysing Rauzy Graphs for DNA Sequences." Thesis, Honours thesis, University of Tasmania, 2005. https://eprints.utas.edu.au/244/1/dnahodil_thesis.pdf.
Full textAbstract:
Bioinformatics is a new and interesting field of research, and genuine discoveries are being made all the time. This project aims to investigate the application of a mathematical tool, called a Rauzy graph, to the problem of classifying DNA sequences. This project aims to learn more about the properties of DNA, and the nature of the graphs themselves.
The work for the project involves developing an algorithm for the computer to generate Rauzy graphs. Different sets of DNA sequences are used to build graphs from, and their properties analysed to investigate their behaviour in the biology domain.
The results of the work show that differences in DNA sequences do affect the structure of their Rauzy graphs and that, through this, the properties of the Rauzy graphs of DNA sequences could be used to classify those sequences.
14
Paleari, Marco. "Informatique Affective : Affichage, Reconnaissance, et Synthèse par Ordinateur des Émotions." Phd thesis, Télécom ParisTech, 2009. http://pastel.archives-ouvertes.fr/pastel-00005615.
Full textAbstract:
L'informatique Affective regarde la computation que se rapporte, surgit de, ou influence délibérément les émotions et trouve son domaine d'application naturel dans les interactions homme-machine a haut niveau d'abstraction. L'informatique affective peut être divisée en trois sujets principaux, à savoir: l'affichage,l'identification, et la synthèse. La construction d'une machine intelligente capable dinteragir'de façon naturelle avec son utilisateur passe forcement par ce trois phases. Dans cette thèse nous proposions une architecture basée principalement sur le modèle dite "Multimodal Affective User Interface" de Lisetti et la théorie psychologique des émotions nommé "Component Process Theory" de Scherer. Dans nos travaux nous avons donc recherché des techniques pour l'extraction automatique et en temps-réel des émotions par moyen des expressions faciales et de la prosodie vocale. Nous avons aussi traité les problématiques inhérentes la génération d'expressions sur de différentes plateformes, soit elles des agents virtuel ou robotique. Finalement, nous avons proposé et développé une architecture pour des agents intelligents capable de simuler le processus humaine d'évaluation des émotions comme décrit par Scherer.
15
Irie, Kenji. "Noise-limited scene-change detection in images." Diss., Lincoln University, 2009. http://hdl.handle.net/10182/1351.
Full textAbstract:
This thesis describes the theoretical, experimental, and practical aspects of a noise-limited method for scene-change detection in images. The research is divided into three sections: noise analysis and modelling, dual illumination scene-change modelling, and integration of noise into the scene-change model. The sources of noise within commercially available digital cameras are described, with a new model for image noise derived for charge-coupled device (CCD) cameras. The model is validated experimentally through the development of techniques that allow the individual noise components to be measured from the analysis of output images alone. A generic model for complementary metal-oxide-semiconductor (CMOS) cameras is also derived. Methods for the analysis of spatial (inter-pixel) and temporal (intra-pixel) noise are developed. These are used subsequently to investigate the effects of environmental temperature on camera noise. Based on the cameras tested, the results show that the CCD camera noise response to variation in environmental temperature is complex whereas the CMOS camera response simply increases monotonically. A new concept for scene-change detection is proposed based upon a dual illumination concept where both direct and ambient illumination sources are present in an environment, such as that which occurs in natural outdoor scenes with direct sunlight and ambient skylight. The transition of pixel colour from the combined direct and ambient illuminants to the ambient illuminant only is modelled. A method for shadow-free scene-change is then developed that predicts a pixel's colour when the area in the scene is subjected to ambient illumination only, allowing pixel change to be distinguished as either being due to a cast shadow or due to a genuine change in the scene. Experiments on images captured in controlled lighting demonstrate 91% of scene-change and 83% of cast shadows are correctly determined from analysis of pixel colour change alone. A statistical method for detecting shadow-free scene-change is developed. This is achieved by bounding the dual illumination model by the confidence interval associated with the pixel's noise. Three benefits arise from the integration of noise into the scene-change detection method: - The necessity for pre-filtering images for noise is removed; - All empirical thresholds are removed; and - Performance is improved. The noise-limited scene-change detection algorithm correctly classifies 93% of scene-change and 87% of cast shadows from pixel colour change alone. When simple post-analysis size-filtering is applied both these figures increase to 95%.
16
Cooke, HBM. "Applying Expert System Technology in a Distributed Monitoring System: An Efficient Response Mechanism for Unknown Threats." Thesis, Honours thesis, University of Tasmania, 2005. https://eprints.utas.edu.au/245/1/Final_Thesis.pdf.
Full textAbstract:
Detecting unknown threats is a paradox; how do you detect a threat if it is not known to exist? The answer is that unknown threat detection is the process of making a previously unknown threat identifiable in the shortest possible time frame. This thesis examines the possibility of creating an unknown threat detection mechanism that security experts can use for developing a flexible protection system for networks. A system that allows the detection of unknown threats through distributed host based monitoring and the incorporation of dynamic and flexible logics with situational knowledge is described as well as the mechanisms used to develop such a system is illustrated. The system not only allows the detection of new threats but does so in a fast and efficient manner to increase the available time for responding to these threats.
17
Bishell, Aaron. "Designing application-specific processors for image processing : a thesis presented in partial fulfilment of the requirements for the degree of Master of Science in Computer Science, Massey University, Palmerston North, New Zealand." Massey University, 2008. http://hdl.handle.net/10179/1024.
Full textAbstract:
Implementing a real-time image-processing algorithm on a serial processor is difficult to achieve because such a processor cannot cope with the volume of data in the low-level operations. However, a parallel implementation, required to meet timing constraints for the low-level operations, results in low resource utilisation when implementing the high-level operations. These factors suggested a combination of parallel hardware, for the low-level operations, and a serial processor, for the high-level operations, for implementing a high-level image-processing algorithm. Several types of serial processors were available. A general-purpose processor requires an extensive instruction set to be able to execute any arbitrary algorithm resulting in a relatively complex instruction decoder and possibly extra FUs. An application-specific processor, which was considered in this research, implements enough FUs to execute a given algorithm and implements a simpler, and more efficient, instruction decoder. In addition, an algorithms behaviour on a processor could be represented in either hardware (i.e. hardwired logic), which limits the ability to modify the algorithm behaviour of a processor, or “software” (i.e. programmable logic), which enables external sources to specify the algorithm behaviour. This research investigated hardware- and software- controlled application-specific serial processors for the implementation of high-level image-processing algorithms and compared these against parallel hardware and general-purpose serial processors. It was found that application-specific processors are easily able to meet the timing constraints imposed by real-time high-level image processing. In addition, the software-controlled processors had additional flexibility, a performance penalty of 9.9% and 36.9% and inconclusive footprint savings (and costs) when compared to hardwarecontrolled processors.
18
Uren, PJ. "Investigation of Distributed and Parallel Performance of a Genetic Algorithm." Thesis, Honours thesis, University of Tasmania, 2004. https://eprints.utas.edu.au/101/1/Philip_Uren_Thesis_Final.pdf.
Full textAbstract:
Genetic algorithms, a stochastic evolutionary computing technique, have demonstrated a capacity for robust, efficient problem solving through highly parallel search space exploration. This work demonstrates how an improvement in performance and efficiency over the traditional serial approach can be achieved by exploiting this highly parallel nature to produce parallel genetic algorithms. Furthermore, it is shown that by incorporating domain specific knowledge into a genetic algorithm near optimal solutions can be located in minimal time.
19
Dazeley, R. "Investigations into Playing Chess Endgames using Reinforcement Learning." Thesis, Honours thesis, University of Tasmania, 2001. https://eprints.utas.edu.au/62/1/Final_Thesis.pdf.
Full textAbstract:
Research in computer game playing has relied primarily on brute force searching approaches rather than any formal AI method. However, these methods may not be able to exceed human ability, as they need human expert knowledge to perform as well as they do. One recently popularized field of research known as reinforcement learning has shown good prospects in overcoming these limitations when applied to non-deterministic games.
This thesis investigated whether the TD(_) algorithm, one method of reinforcement learning, using standard back-propagation neural networks for function generalization, could successfully learn a deterministic game such as chess. The aim is to determine if an agent using no external knowledge can learn to defeat a random player consistently.
The results of this thesis suggests that, even though the agents faced a highly information sparse environment, an agent using a well selected view of the state information was still able to learn to not only to differentiate between various terminating board positions but also to improve its play against a random player. This shows that the reinforcement learning techniques are quite capable of learning behaviour in large deterministic environments without needing any external knowledge.
20
Livingston, JR. "Transfer Of Learnt Knowledge With Card Games." Thesis, Honours thesis, University of Tasmania, 2005. https://eprints.utas.edu.au/246/1/jrl_thesis.pdf.
Full textAbstract:
Reinforcement learning algorithms are an important machine learning technique, which can be applied to the process of learning many tasks.
Much of the existing work on improving these algorithms, and analysis into the usefulness, only considers agents which have to perform one task. Many real-world applications of reinforcement learning algorithms require that an agent can cope with small variations in their given task, and the application of their learnt knowledge to those tasks.
I consider the application of reinforcement learning algorithms to several card games, and the process of transferring learnt knowledge between these card games. The two card games used, Cut-Throat Euchre and Sergent-Major, are similar in their rules and the strategies that are used to play the game. The differences between the two games are used to measure the effectiveness of transferring knowledge between them, using a common-state approach.
These simulations of playing card-games indicate that the tasks of playing these two games are similar enough that knowledge can effectively be shared between the two.
An improvement in the ability of an agent to play one of the games, results in a significant improvement in the ability of the agent to play the other game.
21
Xilan, Tian. "Apprentissage et Noyau pour les Interfaces Cerveau-machine." Phd thesis, INSA de Rouen, 2012. http://tel.archives-ouvertes.fr/tel-00699659.
Full textAbstract:
Les Interface Cerveau-Machine (ICM) ont appliquées avec succès aussi bien dans le domain clinique que pour l'amélioration de la vie quotidienne de patients avec des handicaps. En tant que composante essentielle, le module de traitement du signal détermine nettement la performance d'un système ICM. Nous nous consacrons à améliorer les stratégies de traitement du signal du point de vue de l'apprentissage de la machine. Tout d'abord, nous avons développé un algorithme basé sur les SVM transductifs couplés aux noyaux multiples afin d'intégrer différentes vues des données (vue statistique ou une vue géométrique) dans le processus d'apprentissage. Deuxièmement, nous avons proposé une version en ligne de l'apprentissage multi-noyaux dans le cas supervisé. Les résultats expérimentaux montrent de meilleures performances par rapport aux approches classiques. De plus, l'algorithme proposé permet de sélectionner automatiquement les canaux de signaux EEG utiles grâce à l'apprentissage multi-noyaux. Dans la dernière partie, nous nous sommes attaqué à l'amélioration du module de traitement du signal au-delà des algorithmes d'apprentissage auomatique eux-mêmes. En analysant les données ICM hors-ligne, nous avons d'abord confirmé qu'un modèle de classification simple peut également obtenir des performances statisfaisantes en effectuant une sélection de caractéristiques (et.ou de canaux). Nous avons ensuite conçu un système émotionnel ICM par en tenant compte de l'état émotionnel de l'utilisateur. Sur la base des données de l'EEG obtenus avec différents états émotionnels, c'est-à-dire, positives, négatives et neutre émotions, nous avons finalement prouvé que l'émotion affecter les performances ICM en utilisant des tests statisques. Cette partie de la thèse propose des bases pour réaliser des ICM plus adaptées aux utilisateurs.
22
Manoonpong, Poramate. "Neural preprocessing and control of reactive walking machines : towards versatile artificial perception-action systems /." Berlin : Springer, 2007. http://sfx.ethz.ch/sfx_locater?sid=ALEPH:EBI01&genre=book&isbn=978-3-540-68802-0&id=doi:10.1007/978-3-540-68803-7.
Full text
23
Sun, Zhibin. "Application of artificial neural networks in early detection of Mastitis from improved data collected on-line by robotic milking stations." Lincoln University, 2008. http://hdl.handle.net/10182/665.
Full textAbstract:
Two types of artificial neural networks, Multilayer Perceptron (MLP) and Self-organizing Feature Map (SOM), were employed to detect mastitis for robotic milking stations using the preprocessed data relating to the electrical conductivity and milk yield. The SOM was developed to classify the health status into three categories: healthy, moderately ill and severely ill. The clustering results were successfully evaluated and validated by using statistical techniques such as K-means clustering, ANOVA and Least Significant Difference. The result shows that the SOM could be used in the robotic milking stations as a detection model for mastitis. For developing MLP models, a new mastitis definition based on higher EC and lower quarter yield was created and Principle Components Analysis technique was adopted for addressing the problem of multi-colinearity existed in the data. Four MLPs with four combined datasets were developed and the results manifested that the PCA-based MLP model is superior to other non-PCA-based models in many respects such as less complexity, higher predictive accuracy. The overall correct classification rate (CCR), sensitivity and specificity of the model was 90.74 %, 86.90 and 91.36, respectively. We conclude that the PCA-based model developed here can improve the accuracy of prediction of mastitis by robotic milking stations.
24
Liu, MingHui. "Navel orange blemish identification for quality grading system : a thesis submitted in partial fulfilment of the requirements for the degree of Master of Computer Science at Massey University, Albany, New Zealand." Massey University, 2009. http://hdl.handle.net/10179/1175.
Full textAbstract:
Each year, the world’s top orange producers output millions of oranges for human consumption. This production is projected to grow by as much as 64 million in 2010 and so the demand for fast, low-cost and precise automated orange fruit grading systems is only deemed to become more increasingly important. There is however an underlying limit to most orange blemish detection algorithms. Most existing statistical-based, structural-based, model-based and transform-based orange blemish detection algorithms are plagued by the following problem: any pixels in an image of an orange having about the same magnitudes for the red, green and blue channels will almost always be classified as belonging to the same category (either a blemish or not). This however presents a big problem as the RGB components of the pixels corresponding to blemishes are very similar to pixels near the boundary of an orange. In light of this problem, this research utilizes a priori knowledge of the local intensity variations observed on rounded convex objects to classify the ambiguous pixels correctly. The algorithm has the effect of peeling-off layers of the orange skin according to gradations of the intensity. Therefore, any abrupt discontinuities detected along successive layers would significantly help identifying skin blemishes more accurately. A commercial-grade fruit inspection and distribution system was used to collect 170 navel orange images. Of these images, 100 were manually classified as good oranges by human inspection and the rest are blemished ones. We demonstrate the efficacy of the algorithm using these images as the benchmarking test set. Our results show that the system garnered 96% correctly classified good oranges and 97% correctly classified blemished oranges. The proposed system is easily customizable as it does not require any training. The fruit quality bands can be adjusted to meet the requirements set by the market standards by specifying an agreeable percentage of blemishes for each band.
25
Creemers, Warren. "On the Recognition of Emotion from Physiological Data." Thesis, Edith Cowan University, Research Online, Perth, Western Australia, 2013. https://ro.ecu.edu.au/theses/680.
Full textAbstract:
This work encompasses several objectives, but is primarily concerned with an experiment where 33 participants were shown 32 slides in order to create ‗weakly induced emotions‘. Recordings of the participants‘ physiological state were taken as well as a self report of their emotional state. We then used an assortment of classifiers to predict emotional state from the recorded physiological signals, a process known as Physiological Pattern Recognition (PPR). We investigated techniques for recording, processing and extracting features from six different physiological signals: Electrocardiogram (ECG), Blood Volume Pulse (BVP), Galvanic Skin Response (GSR), Electromyography (EMG), for the corrugator muscle, skin temperature for the finger and respiratory rate. Improvements to the state of PPR emotion detection were made by allowing for 9 different weakly induced emotional states to be detected at nearly 65% accuracy. This is an improvement in the number of states readily detectable. The work presents many investigations into numerical feature extraction from physiological signals and has a chapter dedicated to collating and trialing facial electromyography techniques. There is also a hardware device we created to collect participant self reported emotional states which showed several improvements to experimental procedure.
26
Cont, Arshia. "Traitement et programmation temps-réel des signaux musicaux." Habilitation à diriger des recherches, Université Pierre et Marie Curie - Paris VI, 2013. http://tel.archives-ouvertes.fr/tel-00829771.
Full textAbstract:
Mes activités de recherche et d'encadrement, depuis ma soutenance de thèse en octobre 2008, gravitent à la confluence de deux problématiques souvent considérées comme distinctes dont le couplage est plus qu'évident en informatique musicale : l'écoute artificielle et la programmation synchrone des signaux musicaux en temps réel. Le but est d'enrichir chaque domaine respectivement et de fournir des outils nécessaires aux artistes pour étendre leurs capacités de création musicale sur ordinateur (à celui de leurs homologues humains). Le premier aspect de tout système de musique interactive est l'acte de l'écoute ou formellement parlant, l'extraction d'informations musicaux en temps réel. J'ai concentré mes efforts sur les approches axées sur les applications telles que la détection de pitch polyphonique et les techniques d'alignement en temps réel d'une part. Puis, d'autre part, j'ai abordé des problèmes fondamentaux liés à la représentation des signaux. Pour le premier volet, nous avons proposé deux approches reconnues comme l'état de l'art et renommées par leur utilisation par la communauté artistique et leur reconnaissance au sein de la communauté MIR. Le deuxième volet théorique porte sur le problème fondamental de la quantification et la qualification du contenu de l'information arrivant progressivement dans un système en utilisant des méthodes de la géométrie de l'information. Le deuxième et plus récent aspect de mon travail se concentre sur la programmation synchrone réactive de la musique, couplée à des systèmes d'écoute (ou la formalisation de réactions dans les systèmes interactifs aux perceptions artificielles en temps réel). Notre approche est motivée par les pratiques actuelles en art numérique et les exigences d'évaluation en temps réel qu'elles demandent. L'incarnation majeure de cette approche est le système Antescofo doté d'une machine d'écoute et d'un langage dynamique et qui s'est imposé dans le monde entier dans le répertoire de la musique mixte, reconnu du public depuis 2009. L'écriture du temps, l'interaction et la tentative d'assurer leur exécution correcte en temps réel posent des problèmes difficiles pour l'informatique et les concepteurs de systèmes. Ce couplage fort fera l'objet de mon travail dans les années à venir sous la bannière des systèmes de musique Cyber-physique. Le couplage entre la machine d'écoute et des langages temps réel nécessite d'augmenter les approches actuelles avec des modèles formels de temps, dépassant les approches fonctionnelles qui dominent pour développer une formalisation des exigences de réaction et d'exécution. Il faudra en conséquence repenser nos méthodes usuelles de traitement audio multiforme et leurs modèles de calculs sous-jacents en ce qui concerne le temps.
27
Drira, Hassen. "Calcul statistique sur les variétés de forme pour la l'analyse et la reconnaissance de visage 3D." Phd thesis, Université des Sciences et Technologie de Lille - Lille I, 2011. http://tel.archives-ouvertes.fr/tel-00728009.
Full textAbstract:
Dans cette thèse, nous proposons un cadre Riemannien pour comparer, déformer, calculer des statistiques et organiser de manière hiérarchique des surfaces faciales. Nous appliquons ce cadre à la biométrie faciale 3D où les défis sont les expressions faciales, les variations de la pose et les occultations du visage par des objets externes. Les surfaces faciales sont repr'esentées par un ensemble de courbes de niveaux et de courbes radiales. L'ensemble des courbes fermées (de niveau) constitue une sous-variété non-linéaire de dimension infinie et est utilisé pour représenter le nez, la partie la plus stable du visage. La surface faciale est présentée, par ailleurs, par une collection indexée de courbes radiales. Dans ce cas, le calcul se simplifie et l'espace des formes des courbes ouvertes se ramène à une hyper sphère de l'espace de Hilbert. La comparaison dans l'espace des formes se fait via une métrique élastique afin de faire face aux d'eformations non-isométriques (ne conservant pas les longueurs) des surfaces faciales. Nous proposons des algorithmes pour calculer les moyennes, les vecteurs propres dans ces variétés non-linéaires et l'estimation des parties manquantes des surfaces faciales 3D. L'approche présentée dans cette thèse a été validée sur des Benchmarks connus (FRGCv2, GAVAB, BOSPHORUS) et obtenu des résultats compétitifs par rapport aux méthodes de l'état de l'art.
28
Musti, Utpala. "Synthèse Acoustico-Visuelle de la Parole par Séléction d'Unités Bimodales." Phd thesis, Université de Lorraine, 2013. http://tel.archives-ouvertes.fr/tel-00927121.
Full textAbstract:
Ce travail porte sur la synthèse de la parole audio-visuelle. Dans la littérature disponible dans ce domaine, la plupart des approches traite le problème en le divisant en deux problèmes de synthèse. Le premier est la synthèse de la parole acoustique et l'autre étant la génération d'animation faciale correspondante. Mais, cela ne garantit pas une parfaite synchronisation et cohérence de la parole audio-visuelle. Pour pallier implicitement l'inconvénient ci-dessus, nous avons proposé une approche de synthèse de la parole acoustique-visuelle par la sélection naturelle des unités synchrones bimodales. La synthèse est basée sur le modèle de sélection d'unité classique. L'idée principale derrière cette technique de synthèse est de garder l'association naturelle entre la modalité acoustique et visuelle intacte. Nous décrivons la technique d'acquisition de corpus audio-visuelle et la préparation de la base de données pour notre système. Nous présentons une vue d'ensemble de notre système et nous détaillons les différents aspects de la sélection d'unités bimodales qui ont besoin d'être optimisées pour une bonne synthèse. L'objectif principal de ce travail est de synthétiser la dynamique de la parole plutôt qu'une tête parlante complète. Nous décrivons les caractéristiques visuelles cibles que nous avons conçues. Nous avons ensuite présenté un algorithme de pondération de la fonction cible. Cet algorithme que nous avons développé effectue une pondération de la fonction cible et l'élimination de fonctionnalités redondantes de manière itérative. Elle est basée sur la comparaison des classements de coûts cible et en se basant sur une distance calculée à partir des signaux de parole acoustiques et visuels dans le corpus. Enfin, nous présentons l'évaluation perceptive et subjective du système de synthèse final. Les résultats montrent que nous avons atteint l'objectif de synthétiser la dynamique de la parole raisonnablement bien.
29
Bowie, Mike H. "Evaluation of image analysis for studing mite behaviour." Lincoln University, 1996. http://hdl.handle.net/10182/1056.
Full textAbstract:
The aim of this study was to investigate the usefulness of image analysis for studying mite behaviour. Image analysis was used to convert video recordings of mites' locomotory behaviour into a series of x,y coordinates that, when joined, closely resemble the paths of mites. The coordinates were also used to calculate walking speed, direction of travel, turning frequency, turn bias and tortuosity. Two experimental arenas were developed and used to study the movement of three mite species: 1) a leaf disc arena for two-spotted spider mite, Tetranychus urticae Koch and European red mite, Panonychus ulmi (Koch); and 2), a cover-slip/tack-trap arena for Typhlodromus pyri Scheuten. Two-spotted spider mite exhibited a change in locomotory behaviour through a 48 minute period. Mites exhibited a significant decline in distance travelled, whereas the mean stationary time (per four minute interval) more than doubled, and the duration of stationary events increased steadily over the same period. A reduction in sampling frequency of mite coordinates from one per second to one every two seconds and every four seconds produced a 5% and 12% 'loss' in path length respectively. Sample period length was shown to greatly influence the results produced for some of the mean parameters calculated, however, a reduction in sample length from 3000 to 1500 coordinates was not considered to cause a major loss in information. The influence of the inherent mite movement could not be ignored and made it difficult to make decisions on the 'best' sample length to use. Some strong correlations were found between parameters used to analyse mite locomotory behaviour. In particular, arithmetic mean vector length, speed, total stationary time and total distance travelled were significantly correlated with each other. Mean angular deviation and weighted mean vector length, which both measure the degree of clustering around the mean heading angle, were strongly negatively correlated. Parameters which differentiated between 'straight' and 'tortuous' mite movement were found to be mean meander, absolute mean turn and fractal dimensions. Mean meander was thought to be the most 'powerful', while coefficient of a straight line, a commonly used parameter for measuring tortuosity, did not significantly differentiate between the two different behaviours. Frequency distributions of turns and standard deviations of the three mite species were very similar. All three species had a slight bias to turning right (clockwise) rather than to the left (counter-clockwise) and for each species certain angles occurred more often than would be expected in a 'perfect' normal distribution. A similar pattern also occurred with the frequency distribution of two-spotted spider mite heading angles, in that angles which were expected to occur more often, did not, and vice versa. The potential to use saturated salt solutions to control relative humidity on the arena was` demonstrated and indicated that relative humidity is likely to have an important influence on mite behaviour. Two-spotted spider mites appeared to move more quickly in an attempt to escape the unfavourable, extreme (10% and 95% R.H. at 25°C) moisture conditions. All three mite species displayed a characteristic edge-walking behaviour around the arenas. However, when 'edge' and 'non-edge' behaviours were compared, mean meander was the only parameter (of the parameters tested) which gave a significant difference. Behavioural responses of European red mite and T. pyri to sub-lethal (field rate) esfenvalerate were investigated and the results indicated that these mites did not seek the unsprayed halves of the arenas during the first 48 minutes. However, significant differences in most behavioural parameters to esfenvalerate residues were found with European red mite when whole arenas were compared. Image analysis is an extremely useful research tool for studying mite behaviour because of its ability to measure many parameters quickly. Careful choice of the environmental conditions, the sampling framework, and interpretation of data is essential for meaningful results.
30
Tian, Yuan. "Simulation for LEGO Mindstorms robotics." Lincoln University, 2008. http://hdl.handle.net/10182/304.
Full textAbstract:
The LEGO® MINDSTORMS® toolkit can be used to help students learn basic programming and engineering concepts. Software that is widely used with LEGO MINDSTORMS is ROBOLAB, developed by Professor Chris Rogers from Tufts University, Boston, United States. It has been adopted in about 10,000 schools in the United States and other countries. It is used to program LEGO MINDSTORMS robotics in its icon-based programming environment. However, this software does not provide debug features for LEGO MINDSTORMS programs. Users cannot test the program before downloading it into LEGO robotics hardware. In this project, we develop a simulator for LEGO MINDSTORMS to simulate the motions of LEGO robotics in a virtual 3D environment. We use ODE (Open Dynamic Engine) and OpenGL, combined with ROBOLAB. The simulator allows users to test their ROBOLAB program before downloading it into the LEGO MINDSTORMS hardware. For users who do not have the hardware, they may use the simulator to learn ROBOLAB programming skills which may be tested and debugged using the simulator. The simulator can track and display program execution as the simulation runs. This helps users to learn and understand basic robotics programming concepts. An introduction to the overall structure and architecture of the simulator is given and is followed by a detailed description of each component in the system. This presents the techniques that are used to implement each feature of the simulator. The discussions based on several test results are then given. This leads to the conclusion that the simulator is able to accurately represent the actions of robots under certain assumptions and conditions.
31
Ghazouani, Haythem. "Navigation visuelle de robots mobiles dans un environnement d'intérieur." Phd thesis, Université Montpellier II - Sciences et Techniques du Languedoc, 2012. http://tel.archives-ouvertes.fr/tel-00932829.
Full textAbstract:
Les travaux présentés dans cette thèse concernent le thème des fonctionnalités visuelles qu'il convient d'embarquer sur un robot mobile, afin qu'il puisse se déplacer dans son environnement. Plus précisément, ils ont trait aux méthodes de perception par vision stéréoscopique dense, de modélisation de l'environnement par grille d'occupation, et de suivi visuel d'objets, pour la navigation autonome d'un robot mobile dans un environnement d'intérieur. Il nous semble important que les méthodes de perception visuelle soient à la fois robustes et rapide. Alors que dans les travaux réalisés, on trouve les méthodes globales de mise en correspondance qui sont connues pour leur robustesse mais moins pour être employées dans les applications temps réel et les méthodes locales qui sont les plus adaptées au temps réel tout en manquant de précision. Pour cela, ce travail essaye de trouver un compromis entre robustesse et temps réel en présentant une méthode semi-locale, qui repose sur la définition des distributions de possibilités basées sur une formalisation floue des contraintes stéréoscopiques. Il nous semble aussi important qu'un robot puisse modéliser au mieux son environnement. Une modélisation fidèle à la réalité doit prendre en compte l'imprécision et l'incertitude. Ce travail présente une modélisation de l'environnement par grille d'occupation qui repose sur l'imprécision du capteur stéréoscopique. La mise à jour du modèle est basée aussi sur la définition de valeurs de crédibilité pour les mesures prises. Enfin, la perception et la modélisation de l'environnement ne sont pas des buts en soi mais des outils pour le robot pour assurer des tâches de haut niveau. Ce travail traite du suivi visuel d'un objet mobile comme tâche de haut niveau.
32
Brodie, Matthew Andrew Dalhousie. "Development of fusion motion capture for optimisation of performance in alpine ski racing : a thesis presented in fulfilment of the requirements for the degree of Doctor of Philosophy in Science at Massey University, Wellington, New Zealand." Massey University, 2009. http://hdl.handle.net/10179/1041.
Full textAbstract:
Fusion Motion Capture (FMC), a wearable motion capture system was developed, and applied to the optimisation of athlete performance in alpine ski racing. In what may be a world first, the three-dimensional movements of a skilled athlete (with less than 20 FIS1 points) skiing through a complete training giant slalom racecourse were analysed. FMC consists of multiple light weight sensors attached to the athlete including inertial measurement units (IMUs), pressure sensitive insoles and a global position system (GPS) receiver. The IMUs contain accelerometers, gyroscopes, and magnetometers. Limb orientation and location are obtained by mathematically combining the most reliable data from each sensor using fusion algorithms developed by the author. FMC fuses the signals from the IMUs and GPS without the need for the post filtering, usually applied to motion capture data, and therefore, maintains maximum bandwidth. The FMC results were stable and relatively independent of motion type and duration unlike other inertial systems available in 2005, when the research was initiated. Analysis of data collected from an athlete skiing giant slalom contradict the traditional „going straight turning short? race strategy. The shortest path may not always be the fastest. Instead each gate has a different optimum approach arc. Optimum turn radius increases with both increasing speed and increasing terrain slope. The results also contradict laboratory measurements of ski/snow sliding friction and suggest that snow resistance in giant slalom is of similar importance to wind drag. In addition to gravity, the athlete increased speed using the techniques of „lateral projection? and „pumping?. Race performance was determined from the analysis of the athlete skiing through the entire course. FMC proved, therefore, to be more suitable than traditional optical systems that are practically limited to capturing small sections of a race course. The athlete experienced high and rapidly fluctuating torques about all three axes of the lower joints. This information could be useful in designing training programmes racecourses and equipment to reduce knee injuries. Data driven animations and colour coded force vector diagrams were developed to enhance athlete feedback. Inline skating data was also analysed.
33
Sanderson, Conrad. "Automatic Person Verification Using Speech and Face Information." Thesis, Griffith University, 2003. http://hdl.handle.net/10072/367191.
Full textAbstract:
Identity verification systems are an important part of our every day life. A typical example is the Automatic Teller Machine (ATM) which employs a simple identity verification scheme: the user is asked to enter their secret password after inserting their ATM card; if the password matches the one prescribed to the card, the user is allowed access to their bank account. This scheme suffers from a major drawback: only the validity of the combination of a certain possession (the ATM card) and certain knowledge (the password) is verified. The ATM card can be lost or stolen, and the password can be compromised. Thus new verification methods have emerged, where the password has either been replaced by, or used in addition to, biometrics such as the person’s speech, face image or fingerprints. Apart from the ATM example described above, biometrics can be applied to other areas, such as telephone & internet based banking, airline reservations & check-in, as well as forensic work and law enforcement applications. Biometric systems based on face images and/or speech signals have been shown to be quite effective. However, their performance easily degrades in the presence of a mismatch between training and testing conditions. For speech based systems this is usually in the form of channel distortion and/or ambient noise; for face based systems it can be in the form of a change in the illumination direction. A system which uses more than one biometric at the same time is known as a multi-modal verification system; it is often comprised of several modality experts and a decision stage. Since a multi-modal system uses complimentary discriminative information, lower error rates can be achieved; moreover, such a system can also be more robust, since the contribution of the modality affected by environmental conditions can be decreased. This thesis makes several contributions aimed at increasing the robustness of single- and multi-modal verification systems. Some of the major contributions are listed below. The robustness of a speech based system to ambient noise is increased by using Maximum Auto-Correlation Value (MACV) features, which utilize information from the source part of the speech signal. A new facial feature extraction technique is proposed (termed DCT-mod2), which utilizes polynomial coefficients derived from 2D Discrete Cosine Transform (DCT) coefficients of spatially neighbouring blocks. The DCT-mod2 features are shown to be robust to an illumination direction change as well as being over 80 times quicker to compute than 2D Gabor wavelet derived features. The fragility of Principal Component Analysis (PCA) derived features to an illumination direction change is solved by introducing a pre-processing step utilizing the DCT-mod2 feature extraction. We show that the enhanced PCA technique retains all the positive aspects of traditional PCA (that is, robustness to compression artefacts and white Gaussian noise) while also being robust to the illumination direction change. Several new methods, for use in fusion of speech and face information under noisy conditions, are proposed; these include a weight adjustment procedure, which explicitly measures the quality of the speech signal, and a decision stage comprised of a structurally noise resistant piece-wise linear classifier, which attempts to minimize the effects of noisy conditions via structural constraints on the decision boundary.Thesis (PhD Doctorate)
Doctor of Philosophy (PhD)
School of Microelectronic Engineering
Full Text
34
Susnjak, Teo. "Accelerating classifier training using AdaBoost within cascades of boosted ensembles : a thesis presented in partial fulfillment of the requirements for the degree of Master of Science in Computer Sciences at Massey University, Auckland, New Zealand." Massey University, 2009. http://hdl.handle.net/10179/1002.
Full textAbstract:
This thesis seeks to address current problems encountered when training classifiers within the framework of cascades of boosted ensembles (CoBE). At present, a signifi- cant challenge facing this framework are inordinate classifier training runtimes. In some cases, it can take days or weeks (Viola and Jones, 2004; Verschae et al., 2008) to train a classifier. The protracted training runtimes are an obstacle to the wider use of this framework (Brubaker et al., 2006). They also hinder the process of producing effective object detection applications and make the testing of new theories and algorithms, as well as verifications of others research, a considerable challenge (McCane and Novins, 2003). An additional shortcoming of the CoBE framework is its limited ability to train clas- sifiers incrementally. Presently, the most reliable method of integrating new dataset in- formation into an existing classifier, is to re-train a classifier from beginning using the combined new and old datasets. This process is inefficient. It lacks scalability and dis- cards valuable information learned in previous training. To deal with these challenges, this thesis extends on the research by Barczak et al. (2008), and presents alternative CoBE frameworks for training classifiers. The alterna- tive frameworks reduce training runtimes by an order of magnitude over common CoBE frameworks and introduce additional tractability to the process. They achieve this, while preserving the generalization ability of their classifiers. This research also introduces a new framework for incrementally training CoBE clas- sifiers and shows how this can be done without re-training classifiers from beginning. However, the incremental framework for CoBEs has some limitations. Although it is able to improve the positive detection rates of existing classifiers, currently it is unable to lower their false detection rates.
35
Yasin, Seyed Mohammad Taghi Alhoseyni Almodarresi. "Application of intelligent signal processing to dynamic measurement systems." Thesis, University of Southampton, 2001. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.342812.
Full text
36
Hassan, K. J. "Application of artificial neural networks for understanding and diagnosing the state of mastitis in dairy cattle." Lincoln University, 2007. http://hdl.handle.net/10182/633.
Full textAbstract:
Bovine mastitis adversely affects the dairy industry around the world. This disease is caused by a diverse range of bacteria, broadly categorised as minor and major pathogens. In-line tools that help identify these bacterial groupings in the early stages of the disease are advantageous as timely decisions could be made before the cow develops any clinical symptoms. The first objective of this research was to identify the most informative milk parameters for the detection of minor and major bacterial pathogens. The second objective of this research was to evaluate the potential of supervised and unsupervised neural network learning paradigms for the detection of minor infected and major infected quarters in the early stages of the disease. The third objective was to evaluate the effects of different proportions of infected to non-infected cases in the training data set on the correct classification rate of the supervised neural network models as there are proportionately more non-infected cases in a herd than infected cases. A database developed at Lincoln University was used to achieve the research objectives. Starting at calving, quarter milk samples were collected weekly from 112 cows for a period of fourteen weeks, resulting in 4852 samples with complete records for somatic cell count (SCC), electrical resistance, protein percentage, fat percentage, and bacteriological status. To account for the effects of the stage of lactation on milk parameters with respect to days in milking, data was divided into three days in milk ranges. In addition, cow variation was accounted for by the sire family from which the cow originated and the lactation number of each cow. Data was pre-processed before the application of advanced analytical techniques. Somatic cell score (SCS) and electrical resistance index were derived from somatic cell count and electrical resistance, respectively. After pre-processing, the data was divided into training and validation sets for the unsupervised neural network modelling experiment and, for the supervised neural network modelling experiments, the data was divided into training, calibration and validation sets. Prior to any modelling experiments, the data was analysed using statistical and multivariate visualisation techniques. Correlations (p<0.05) were found between the infection status of a quarter and its somatic cell score (SCS, 0.86), electrical resistance index (ERI, -0.59) and protein percentage (PP, 0.33). The multivariate parallel visualisation analysis validated the correlation analysis. Due to significant multicolinearity [Correlations: SCS and ERI (-0.65: p<0.05); SCS and PP (0.32: p<0.05); ERI and PP (-0.35: p<0.05)], the original variables were decorrelated using principle component analysis. SCS and ERI were found to be the most informative variables for discriminating between non-infected, minor infected and major infected cases. Unsupervised neural network (USNN) model was trained using the training data set which was extracted from the database, containing approximately equal number of randomly selected records for each bacteriological status [not infected (NI), infected with a major pathogen (MJI) and infected with a minor pathogen (MNI)]. The USNN model was validated with the remaining data using the four principle components, days in milk (DIM), lactation number (LN), sire number, and bacteriological status (BS). The specificity of the USNN model in correctly identifying non infected cases was 97%. Sensitivities for correctly detecting minor and major infections were 89% and 80%, respectively. The supervised neural network (SNN) models were trained, calibrated and validated with several sets of training, calibration and validation data, which were randomly extracted from the database in such a way that each set has a different proportion of infected to non-infected cases ranging from 1:1 to 1:10. The overall accuracy of these models based on validation data sets gradually increased with increase in the number of non-infected cases in the data sets (80% for the 1:1, 84% for 1:2, 86% for 1:4 and 93% for 1:10). Specificities of the best models for correctly recognising non-infected cases for the four data sets were 82% for 1:1, 91% for 1:2, 94% for 1:4 and 98% for 1:10. Sensitivities for correctly recognising minor infected cases for the four data sets were 86% for 1:1, 76% for 1:2, 71% for 1:4 and 44% for 1:10. Sensitivities for correctly recognising major infected cases for the four data sets were 20% for 1:1, 20% for 1:2, 30% for 1:4 and 40% for 1:10. Overall, sensitivity for the minor infected cases decreased while that of major infected cases increased with increase in the number non-infected cases in the training data set. Due to the very low prevalence of MJI category in this particular herd, results for this category may be inconclusive. This research suggests that somatic cell score and electrical resistance index of milk were the most effective variables for detecting the infection status of a quarter followed by milk protein and fat percentage. The neural network models were able to differentiate milk containing minor and major bacterial pathogens based on milk parameters associated with mastitis. It is concluded that the neural network models can be developed and incorporated into milking machines to provide an efficient and effective method for the diagnosis of mastitis.
37
Wong, Hau San. "Adaptive image processing using computational intelligence techniques." Thesis, The University of Sydney, 1998. https://hdl.handle.net/2123/27658.
Full textAbstract:
In this thesis, we illustrate the essential aspects of the adaptive image processing problem
in terms of two applications: the adaptive assignment of the regularization parameters in
image restoration, and the adaptive characterization of edges in feature detection applications.
These two problems are representative of the general adaptive image processing
paradigm in that the three requirements for its successive implementation: namely the
segmentation of an image into its main feature types, the characterization of each of these
features, and the optimization of the image model parameters corresponding to the individual
features, are present. In view of these requirements, we have adopted the three
main approaches within the class of computational intelligence algorithms, namely neu—
ral network techniques, fuzzy set theory, and evolutionary computation, for solving the
adaptive image processing problem. This is in view of the direct correspondence between
some of the above requirements with the particular capabilities of specific computational
intelligence approaches.
We first applied neural network techniques to the adaptive regularization problem in
image restoration. Instead of the usual approach of selecting the regularization parameter
values by trial and error, we adopt a learning approach by treating the parameters in
various local image regions as network weights of a model—based neural network with
hierarchical architecture (HMBNN), such that they are adjustable through the supply of
training examples specifying the desired image quality.
In addition, we also applied the HMBNN to the problem
38
Delaye, Adrien. "Méta-modèles de positionnement spatial pour la reconnaissance de tracés manuscrits." Phd thesis, INSA de Rennes, 2011. http://tel.archives-ouvertes.fr/tel-00651446.
Full textAbstract:
L'essor des interfaces homme-machine permettant la saisie d'informations à l'aide d'un stylo électronique est accompagné par le développement de méthodes automatiques pour interpréter des données de plus en plus riches et complexes : texte manuscrit, mais aussi expressions mathématiques, schémas, prise de notes libre... Pour interpréter efficacement ces documents manuscrits, il est nécessaire de considérer conjointement les formes des objets qui les constituent et leur positionnement spatial. Nos recherches se concentrent sur la modélisation du positionnement spatial entre des objets manuscrits, en partant du constat qu'il n'est pas exploité dans toute sa richesse par les méthodes actuelles. Nous introduisons le concept de méta-modèle spatial, une modélisation générique pour décrire des relations spatiales entre des objets de nature, complexité et formes variables. Ces modèles, qui peuvent être appris à partir de données, offrent une richesse et une précision inédite car ils autorisent la conduite d'un raisonnement spatial directement dans l'espace image. L'appui sur le cadre de la théorie des sous-ensembles flous et de la morphologie mathématique permet la gestion de l'imprécision et offre une description des relations spatiales conforme à l'intuition. Un méta-modèle est doté d'un pouvoir de prédiction qui permet de décrire la relation spatiale modélisée au sein de l'image, par rapport à un objet de référence. Cette capacité rend possible la visualisation des modèles et fournit un outil pour segmenter les tracés en fonction de leur contexte. En exploitant ces modèles, nous proposons une représentation pour des objets manuscrits à la structure complexe. Cette représentation repose uniquement sur la modélisation de leurs informations spatiales, afin de démontrer l'importance de ces informations pour l'interprétation d'objets manuscrits structurés. La segmentation des tracés en primitives structurelles est guidée par les modèles de positionnement, via leur capacité de prédiction. Les résultats expérimentaux, portant sur des objets de complexité et de natures diverses (caractères chinois, gestes d'édition, symboles mathématiques, lettres), confirment la bonne qualité de description du positionnement offerte par les méta-modèles. Les tests de reconnaissance de symboles par l'exploitation de leur information spatiale attestent d'une part de l'importance de cette information et valident d'autre part la capacité des méta-modèles à la représenter avec une grande précision. Ces résultats témoignent donc de la richesse de l'information spatiale et du potentiel des méta-modèles spatiaux pour l'amélioration des techniques de traitement du document manuscrit.
39
Gadde, Pramod. "AFFINE IMAGE REGISTRATION USING ARTIFICIAL NEURAL NETWORKS." DigitalCommons@CalPoly, 2013. https://digitalcommons.calpoly.edu/theses/982.
Full textAbstract:
This thesis deals with image registration of MRI images using neural networks. Image registration combines multiple images of the same subject that were taken at different points in time, from different sensors, or from different points of views into a single image and coordinate system. Image registration is widely used in medical imaging and remote sensing. In this thesis feed forward neural networks and wavelet neural networks are used to estimate the parameters of registration. Simulations show that the wavelet networks provide significantly more accurate results than feed forward networks and other proposed methods including genetic algorithms. Both methods are also shown to be robust to noise and changes in parameter ranges.
40
Mahdjoub, Jason. "Vers un système de vision auto-adaptatif à base de systèmes multi-agents." Phd thesis, Université de Reims - Champagne Ardenne, 2011. http://tel.archives-ouvertes.fr/tel-00937422.
Full textAbstract:
Il existe une multitude de traitements d'images dans la littérature, chacun étant adapté à un ensemble plus ou moins grand de cadres d'application. La généralisation ou la mise en collaboration de ces traitements pour un système plus complet et plus robuste est un problème mal posé. Les traitements d'images sont fondamentalement trop différents les uns par rapport aux autres pour être mis en commun de façon naturelle. De plus, ces derniers sont trop rigides pour pouvoir s'adapter d'eux-mêmes lorsqu'un problème non prévu à l'avance par le concepteur apparaît. Or la vision est un phénomène autoadaptatif, qui sait traiter en temps réel des situations singulières, en y proposant des traitements particuliers et adaptés. Elle est aussi un traitement complexe des informations, tant ces dernières ne peuvent être réduites à des représentations réductionnistes et simplifiantes sans être mutilées. Dans cette thèse, un système de vision est entrepris comme un tout où chaque partie est adaptée à l'autre, mais aussi où chaque partie ne peut s'envisager sans l'autre dans les tensions les plus extrêmes générées par la complexité et l'intrication des informations. Puisque chaque parcelle d'information joue un rôle local dans la vision, tout en étant dirigée par un objectif global peu assimilable à son niveau, nous envisageons la vision comme un système où chaque agent délibère selon une interférence produite par le potentiel décisionnel de chacun de ses voisins. Cette délibération est entreprise comme le résultat produit par l'interférence d'une superposition de solutions. De cette manière, il émerge du système à base d'agents une décision commune qui dirige les actions locales faites par chaque agent ou chaque partie du système. En commençant par décrire les principales méthodes de segmentation ainsi que les descripteurs de formes, puis en introduisant les systèmes multi-agents dans le domaine de l'image, nous discutons d'une telle approche où la vision est envisagée comme un système multi-agent apte à gérer la complexité inhérente de l'information visuelle tant en représentation qu'en dynamisme systémique. Nous ancrons dans ces perspectives deux modèles multi-agents. Le premier modèle traite de la segmentation adaptative d'images sans calibration manuelle par des seuils. Le deuxième modèle traite de la représentation de formes quelconques à travers la recherche de coefficients d'ondelettes pertinents. Ces deux modèles remplissent des critères classiques liés au traitement d'images, et à la reconnaissance de formes, tout en étant des cas d'études à développer pour la recherche d'un système de vision auto-adaptatif tel que nous le décrivons.
41
Ouni, Slim. "Parole Multimodale : de la parole articulatoire à la parole audiovisuelle." Habilitation à diriger des recherches, Université de Lorraine, 2013. http://tel.archives-ouvertes.fr/tel-00927119.
Full textAbstract:
La communication parlée est par essence multimodale. Le signal acoustique véhicule la modalité auditive, et l'image la modalité visuelle et gestuelle (déformations du visage). Le signal de parole est en effet la conséquence des déformations du conduit vocal sous l'effet du mouvement de la mâchoire, des lèvres, de la langue, etc.. pour moduler le signal d'excitation produit par les cordes vocales ou les turbulences aérodynamiques. Ces déformations sont visibles au niveau du visage (lèvres, joues, mâchoire) grâce à la coordination des différents muscles orofaciaux et de la déformation de la peau induite par ces derniers. La modalité visuelle permet de fournir des informations complémentaires au signal acoustique, et elle devient indispensable dans le cas où le signal acoustique est dégradé, comme c'est le cas chez les malentendants, ou en milieu bruité. D'autres modalités peuvent être liées à la parole, comme les mouvements des sourcils et les différents gestes qui expriment l'émotion. Cette dernière modalité suprasegmentale peut, comme la modalité visuelle, compléter le message acoustique ou acoustique-visuel. Cet exposé présentera les travaux que je mène sur la parole multimodale. Ce caractère multimodal de la communication parlée est traité de deux façons différentes : (1) étudier les deux composantes articulatoire et acoustique de la parole. En effet, je m'intéresse à la caractérisation articulatoire des sons et à l'étude du lien entre l'espace articulatoire et l'espace acoustique. En particulier, je m'intéresse à la récupération de l'évolution temporelle du conduit vocal à partir du signal acoustique (aussi appelée inversion acoustique-articulatoire) et à l'étude de la caractérisation articulatoire de la parole par l'analyse de corpus de données articulatoires. (2) étudier les deux composantes acoustique et visuelle. Dans ce cadre, je m'intéresse à l'effet de la déformation du conduit vocal sur l'apparence du visage qui véhicule le message visuel. La synthèse acoustique-visuelle est un cadre qui permet d'étudier cet aspect. De plus, l'étude de l'intelligibilité audiovisuelle permet de mieux comprendre les mécanismes de la communication audiovisuelle, mais également d'évaluer le système de synthèse acoustique-visuelle. Enfin, je présenterai mon programme de recherche qui porte sur la parole multimodale expressive que je propose d'étudier globalement, c.-à-d. en considérant les composantes articulaire, acoustique et visuelle ainsi que l'expressivité intrinsèque de celles-ci, simultanément. Je propose en particulier d'aborder la modélisation de la dynamique articulatoire et faciale de la parole pour produire de la parole combinée avec les expressions faciales.
42
Choi, Hyunjong. "Medical Image Registration Using Artificial Neural Network." DigitalCommons@CalPoly, 2015. https://digitalcommons.calpoly.edu/theses/1523.
Full textAbstract:
Image registration is the transformation of different sets of images into one coordinate system in order to align and overlay multiple images. Image registration is used in many fields such as medical imaging, remote sensing, and computer vision. It is very important in medical research, where multiple images are acquired from different sensors at various points in time. This allows doctors to monitor the effects of treatments on patients in a certain region of interest over time. In this thesis, artificial neural networks with curvelet keypoints are used to estimate the parameters of registration. Simulations show that the curvelet keypoints provide more accurate results than using the Discrete Cosine Transform (DCT) coefficients and Scale Invariant Feature Transform (SIFT) keypoints on rotation and scale parameter estimation.
43
Sinha, Sunil K. "Automated underground pipe inspection using a unified image processing and artificial intelligence methodology." Thesis, National Library of Canada = Bibliothèque nationale du Canada, 2000. http://www.collectionscanada.ca/obj/s4/f2/dsk1/tape4/PQDD_0015/NQ53517.pdf.
Full text
44
Vieira, Fábio Henrique Antunes [UNESP]. "Image processing through machine learning for wood quality classification." Universidade Estadual Paulista (UNESP), 2016. http://hdl.handle.net/11449/142813.
Full textAbstract:
Submitted by FÁBIO HENRIQUE ANTUNES VIEIRA null (curso_structural@hotmail.com) on 2016-08-03T12:43:17Z
No. of bitstreams: 1
Fábio Henrique Antunes Vieira TESE.pdf: 4977174 bytes, checksum: f3e115728925e457e12dd4a79c93812a (MD5)Approved for entry into archive by Ana Paula Grisoto (grisotoana@reitoria.unesp.br) on 2016-08-04T19:15:49Z (GMT) No. of bitstreams: 1 vieira_fha_dr_guara.pdf: 4977174 bytes, checksum: f3e115728925e457e12dd4a79c93812a (MD5)
Made available in DSpace on 2016-08-04T19:15:49Z (GMT). No. of bitstreams: 1 vieira_fha_dr_guara.pdf: 4977174 bytes, checksum: f3e115728925e457e12dd4a79c93812a (MD5) Previous issue date: 2016-06-30
A classificação da qualidade da madeira é indicada para indústria de processamento e produção desse material. Essas empresas têm investido em soluções para agregar valor à matéria-prima, com o intuito de melhorar resultados, observando os rumos do mercado. O objetivo deste trabalho foi comparar Redes Neurais Convolutivas, um método de aprendizado profundo, na classificação da qualidade de madeira, com outras técnicas tradicionais de Máquinas de aprendizado, como Máquina de Vetores de Suporte, Árvores de Decisão, Regra dos Vizinhos Mais Próximos e Redes Neurais, em conjunto com Descritores de Textura. Isso foi possível através da verificação do nível de acurácia das experiências com diferentes técnicas, como Aprendizado Profundo e Descritores de Textura no processamento de imagens destes objetos. Foi utilizada uma câmera convencional para capturar as 374 amostras de imagem adotadas no experimento, e a base de dados está disponível para consulta. O processamento das imagens passou por algumas fases, após terem sido obtidas, como pré-processamento, segmentação, análise de recursos e classificação. Os métodos de classificação se deram através de Aprendizado Profundo e por meio de técnicas de Aprendizado de Máquinas tradicionais como Máquina de Vetores de Suporte, Árvores de Decisão, Regra dos Vizinhos Mais Próximos e Redes Neurais juntamente com os Descritores de Textura. Os resultados empíricos para o conjunto de dados das imagens da madeira serrada mostraram que o método com Descritores de Textura, independentemente da estratégia empregada, foi muito competitivo quando comparado com as Redes Neurais Convolutivas para todos os experimentos realizados, e até mesmo superou-as para esta aplicação.
The quality classification of wood is prescribed throughout the wood chain industry, particularly those from the processing and manufacturing fields. Those organizations have invested energy and time trying to increase value of basic items, with the purpose of accomplishing better results, in agreement to the market. The objective of this work was to compare Convolutional Neural Network, a deep learning method, for wood quality classification to other traditional Machine Learning techniques, namely Support Vector Machine (SVM), Decision Trees (DT), K-Nearest Neighbors (KNN), and Neural Networks (NN) associated with Texture Descriptors. Some of the possible options were to assess the predictive performance through the experiments with different techniques, Deep Learning and Texture Descriptors, for processing images of this material type. A camera was used to capture the 374 image samples adopted on the experiment, and their database is available for consultation. The images had some stages of processing after they have been acquired, as pre-processing, segmentation, feature analysis, and classification. The classification methods occurred through Deep Learning, more specifically Convolutional Neural Networks - CNN, and using Texture Descriptors with Support Vector Machine, Decision Trees, K-nearest Neighbors and Neural Network. Empirical results for the image dataset showed that the approach using texture descriptor method, regardless of the strategy employed, is very competitive when compared with CNN for all performed experiments, and even overcome it for this application.
45
Gooch, Richard M. "Machine learning techniques for signal processing, pattern recognition and knowledge extraction from examples." Thesis, University of Bristol, 1995. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.294898.
Full text
46
Baccouche, Moez. "Apprentissage neuronal de caractéristiques spatio-temporelles pour la classification automatique de séquences vidéo." Phd thesis, INSA de Lyon, 2013. http://tel.archives-ouvertes.fr/tel-00932662.
Full textAbstract:
Cette thèse s'intéresse à la problématique de la classification automatique des séquences vidéo. L'idée est de se démarquer de la méthodologie dominante qui se base sur l'utilisation de caractéristiques conçues manuellement, et de proposer des modèles qui soient les plus génériques possibles et indépendants du domaine. Ceci est fait en automatisant la phase d'extraction des caractéristiques, qui sont dans notre cas générées par apprentissage à partir d'exemples, sans aucune connaissance a priori. Nous nous appuyons pour ce faire sur des travaux existants sur les modèles neuronaux pour la reconnaissance d'objets dans les images fixes, et nous étudions leur extension au cas de la vidéo. Plus concrètement, nous proposons deux modèles d'apprentissage des caractéristiques spatio-temporelles pour la classification vidéo : (i) Un modèle d'apprentissage supervisé profond, qui peut être vu comme une extension des modèles ConvNets au cas de la vidéo, et (ii) Un modèle d'apprentissage non supervisé, qui se base sur un schéma d'auto-encodage, et sur une représentation parcimonieuse sur-complète des données. Outre les originalités liées à chacune de ces deux approches, une contribution supplémentaire de cette thèse est une étude comparative entre plusieurs modèles de classification de séquences parmi les plus populaires de l'état de l'art. Cette étude a été réalisée en se basant sur des caractéristiques manuelles adaptées à la problématique de la reconnaissance d'actions dans les vidéos de football. Ceci a permis d'identifier le modèle de classification le plus performant (un réseau de neurone récurrent bidirectionnel à longue mémoire à court-terme -BLSTM-), et de justifier son utilisation pour le reste des expérimentations. Enfin, afin de valider la généricité des deux modèles proposés, ceux-ci ont été évalués sur deux problématiques différentes, à savoir la reconnaissance d'actions humaines (sur la base KTH), et la reconnaissance d'expressions faciales (sur la base GEMEP-FERA). L'étude des résultats a permis de valider les approches, et de montrer qu'elles obtiennent des performances parmi les meilleures de l'état de l'art (avec 95,83% de bonne reconnaissance pour la base KTH, et 87,57% pour la base GEMEP-FERA).
47
Al-khedher, Mohammad Abdelfatah. "Carbon nanotubes characterization and quality analysis using artificial intelligence." Online access for everyone, 2007. http://www.dissertations.wsu.edu/Dissertations/Spring2007/m_al-khedher_041907.pdf.
Full text
48
Kittel, W. Armin. "A symbolic representation for signals." Diss., Georgia Institute of Technology, 1992. http://hdl.handle.net/1853/15820.
Full text
49
Moshtagh, Jamal. "Accurate fault location in low voltage underground cables using advanced signal processing and artificial intelligence." Thesis, University of Bath, 2006. https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.433963.
Full text
50
Luwes, Nicolaas Johannes. "Artificial intelligence machine vision grading system." Thesis, Bloemfontein : Central University of Technology, Free State, 2014. http://hdl.handle.net/11462/35.
Full text