Dissertations / Theses on the topic 'COMPLEX BACKGROUND'
To see the other types of publications on this topic, follow the link: COMPLEX BACKGROUND.
Create a spot-on reference in APA, MLA, Chicago, Harvard, and other styles
Consult the top 47 dissertations / theses for your research on the topic 'COMPLEX BACKGROUND.'
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
Albalooshi, Fatema A. "Self-organizing Approach to Learn a Level-set Function for Object Segmentation in Complex Background Environments." University of Dayton / OhioLINK, 2015. http://rave.ohiolink.edu/etdc/view?acc_num=dayton1429545327.
Full text
2
Mohapatra, Deepankar. "Automatic Removal of Complex Shadows From Indoor Videos." Thesis, University of North Texas, 2015. https://digital.library.unt.edu/ark:/67531/metadc804942/.
Full textAbstract:
Shadows in indoor scenarios are usually characterized with multiple light sources that produce complex shadow patterns of a single object. Without removing shadow, the foreground object tends to be erroneously segmented. The inconsistent hue and intensity of shadows make automatic removal a challenging task. In this thesis, a dynamic thresholding and transfer learning-based method for removing shadows is proposed. The method suppresses light shadows with a dynamically computed threshold and removes dark shadows using an online learning strategy that is built upon a base classifier trained with manually annotated examples and refined with the automatically identified examples in the new videos. Experimental results demonstrate that despite variation of lighting conditions in videos our proposed method is able to adapt to the videos and remove shadows effectively. The sensitivity of shadow detection changes slightly with different confidence levels used in example selection for classifier retraining and high confidence level usually yields better performance with less retraining iterations.
3
Isús, Díaz Laura 1989. "Creating ad hoc functional ontologies for complex diseases and understanding drug response with cellular networks." Doctoral thesis, Universitat Pompeu Fabra, 2017. http://hdl.handle.net/10803/565810.
Full textAbstract:
Since the advent of the genomic era, vast amounts of data from heterogeneous biomedical sources have been generated, revealing an overwhelming complexity and creating a need for new computational tools. The etiology for human diseases is complex, involving the malfunction of multiple genes in the intricate molecular networks that form the cell machinery. In the same direction, lower levels of complexity, our cells, are also complex and the outcome of an external perturbation cannot be directly predicted from the genotype.
In this thesis, we address biological complexity at different levels of organization. First, we study complex diseases from the manual generation of ad hoc functional ontologies. Then, we decipher cell response to external perturbations revealing the role that the genetic background plays on the final outcome. Unraveling biological complexity requires the integration of heterogeneous biological data, at different levels of organization and the combination of multiple computational strategies.Des de l'origen de l'era genòmica, s'han generat grans quantitats de dades biomèdiques, revelant una enorme complexitat i creant la necessitat de nous mètodes computacionals. L'etiologia de les malalties humanes és complexa, implicant el mal funcionament de múltiples gens i de la intricada xarxa molecular que conforma la maquinària cel·lular. De la mateixa manera, nivells més simples de complexitat, les nostres cèl·lules, són tanmateix complexes i la resposta a una pertorbació externa és difícil de predir a partir del seu genotip. En aquesta tesi, abordem la complexitat biològica a diferents nivells d'organització. Primer, estudiem les malalties complexes des de la generació ad hoc d'ontologies funcionals. Després, analitzem la resposta cel·lular a pertorbacions externes, revelant el paper que juga el rerefons genètic en el fenotip final. Desentranyar la complexitat biològica requereix la integració de dades biològiques heterogènies, a diferents nivells d'organització i la combinació de múltiples estratègies computacionals.
4
Goshovska, A. V. "Features of the vascular component at the stage of the placental complex formation against a background of inflammatory diseases of the female genital organs." Thesis, БДМУ, 2021. http://dspace.bsmu.edu.ua:8080/xmlui/handle/123456789/18715.
Full text
5
GOMES, PABLO FRIAS DE OLIVEIRA BIONE. "FOREGROUND EXTRACTION IN HD IMAGES ON COMPLEX BACKGROUNDS." PONTIFÍCIA UNIVERSIDADE CATÓLICA DO RIO DE JANEIRO, 2010. http://www.maxwell.vrac.puc-rio.br/Busca_etds.php?strSecao=resultado&nrSeq=16308@1.
Full textAbstract:
PONTIFÍCIA UNIVERSIDADE CATÓLICA DO RIO DE JANEIROA indústria de cinema e TV tem usado amplamente a técnica de Chroma Key, também conhecida por Blue Screen Matting. Esta técnica revolucionou, ao longo do tempo, a indústria do entretenimento, permitindo que cenas impossíveis de serem criadas se tornassem realidade. A evolução dessa técnica permitiu que produções complexas passassem a ter melhor controle e custos mais baixos. Porém, essa técnica conta com uma série de etapas de preparação, que demandam recursos financeiros elevados e planejamento preciso. Ademais, erros de continuidade costumam criar sérios problemas na pós-produção. Atualmente, a indústria de entretenimento está procurando outras técnicas de matting que funcionem com fundos variados. O uso destas técnicas ainda está restrita a trabalhos acadêmicos e a softwares de manipulação de imagens estáticas. O presente trabalho tem como objetivos fazer uma análise dos processos atuais de chroma key e partir para a proposta de uma técnica de matting com fundos variados em imagens de alta definição (HD – High Definition). Dois métodos para o cálculo de valores de alpha são apresentados: um método global baseado em clusters e um método local baseado em potencial elétrico.
The film and broadcast industry have been massively using the Chroma Key technique, also known as Blue Screen Matting. This technique deeply transformed the entertainment industry, allowing impossible scenes become reality. The evolution of this technique allowed that complex productions could have better control and lower costs. However, this technique needs a sequence of preparation stages, which require high budgets and precise planning. Furthermore, continuity errors usually cause serious post-production problems. Currently, the entertainment industry is searching for other matting techniques that work on any kind of background. The use of these techniques is still restricted to academic works and softwares of still image manipulation. The present work has the goal of making an analysis of the current chroma key processes and aims to propose a matting technique over any type of background in High Definition images. Two methods of calculating alpha values are presented: a local method based on clusters and a local one based on electric potential.
6
Haywood, Nicholas Russell. "Build-up and resetting of auditory stream segregation in quiet and in complex-tone backgrounds." Thesis, Aston University, 2009. http://publications.aston.ac.uk/15319/.
Full textAbstract:
Thirteen experiments investigated the dynamics of stream segregation. Experiments 1-6b used a similar method, where a same-frequency induction sequence (usually 10 repetitions of an identical pure tone) promoted segregation in a subsequent, briefer test sequence (of alternating low- and high-frequency tones). Experiments 1-2 measured streaming using a direct report of perception and a temporal-discrimination task, respectively. Creating a single deviant by altering the final inducer (e.g. in level or replacement with silence) reduced segregation, often substantially. As the prior inducers remained unaltered, it is proposed that the single change actively reset build-up. The extent of resetting varied gradually with the size of a frequency change, once noticeable (experiments 3a-3b). By manipulating the serial position of a change, experiments 4a-4b demonstrated that resetting only occurred when the final inducer was replaced with silence, as build-up is very rapid during a same-frequency induction sequence. Therefore, the observed resetting cannot be explained by fewer inducers being presented. Experiment 5 showed that resetting caused by a single deviant did not increase when prior inducers were made unpredictable in frequency (four-semitone range). Experiments 6a-6b demonstrated that actual and perceived continuity have a similar effect on subsequent streaming judgements promoting either integration or segregation, depending on listening context. Experiment 7 found that same-frequency inducers were considerably more effective at promoting segregation than an alternating-frequency inducer, and that a trend for deviant-tone resetting was only apparent for the same-frequency case.
7
Loreti, Bianca Angelica. "Validity of a Nonspeech Dynamic Assessment of Phonological Awareness in Children from Spanish-speaking Backgrounds." Scholar Commons, 2015. https://scholarcommons.usf.edu/etd/5732.
Full textAbstract:
Literacy development in Spanish-speaking children is a growing concern in the United States (Invernizzi, 2009). Phonological awareness is a predictor of literacy achievement in most alphabetic languages (Anthony et al., 2011; Davison & Brea-Spahn, 2012; Durgunoğlu, Nagy, & Hancin-Bhatt, 1993; Goikoetxea, 2005). Bilingual children with complex communication needs (CCN) demonstrate increased difficulties in speaking, reading, and writing, making learning two languages a difficult task (Toppelberg, Snow, & Tager-Flusberg, 1999). Literacy attainment in bilingual individuals who have CCN is important to improve their overall language development and communication interaction skills (Harrison-Harris, 2002). A valid and reliable phonological awareness assessment that does not require speech is needed in order to provide appropriate instruction and address desired literacy goals (Barker, Bridges, & Saunders, 2014).
The goal of this study is to describe pilot data from the Dynamic Assessment of Phonemic Awareness in Spanish (DAPA-S), a new dynamic phonological awareness assessment that does not require speech responses, with children from Latin American Spanish-speaking backgrounds, in order to determine its construct validity. DAPA-S was administered over the course of one to three sessions to ten participants (six males and four females). Participants also received the Identificación de letras y palabras (Letter-Word Identification; LWID) subscale from the Woodcock-Muñoz Language Survey–Revised (WMLS-R; Woodcock, Muñoz- Sandoval, Ruef, & Alvarado, 2005) as an emergent reading skill task and three subtests from the Test of Phonological Sensitivity in Spanish (TOPSS; Brea, Silliman, Bahr, & Bryant, 2003):letter-name and letter-sound, elision, and rapid automatized naming (RAN) as assessments of phonological awareness.
To evaluate concurrent validity, Pearson correlations and bootstrapped 95% confidence intervals were calculated between the DAPA-S total score and the measures of phonological awareness from the TOPSS. The DAPA-S demonstrated strong and significant correlations with elision, RAN, and the letter-sound subtests rs = –.67 to .87, ps = .00 to .03. These results indicated that the DAPA-S likely measured the same construct as the other measures of phonological awareness from the TOPSS.
To evaluate convergent validity, Pearson correlations and bootstrapped 95% confidence intervals were calculated between LWID of the WMLS-R and the DAPA-S total score. The DAPA-S demonstrated a strong and significant correlation, r = .75, p < .05.
The data suggest a high degree of both concurrent and convergent validity, as many of the conventional measures of phonological awareness and emergent reading were significantly correlated with the DAPA-S, including letter-sound, RAN, and LWID. Overall, the pattern of results suggests that the DAPA-S may be a reliable and valid tool for measurement of phonological awareness in Spanish.
8
Havlíček, Jakub. "Zázemí sportovního areálu - stavebně technologický projekt." Master's thesis, Vysoké učení technické v Brně. Fakulta stavební, 2020. http://www.nusl.cz/ntk/nusl-409923.
Full textAbstract:
This diploma thesis is dealing with a construction and technological project of the backgrounds and changing rooms of the sports centre in Velké Meziříčí. An engineering report of the construction and technological project, a suggested plan of transport routes and a time and financial plan for all the building facilities are included. With a realization study of the main technological phases, a site equipment project and a suggested list of the main construction vehicles and mechanisms is being dealt in another part of the thesis. In addition to these there is a time plan of the main construction object, an itemized budget, a time deployment of machines and workers and a plan for securing all the material resources needed. An engineering report for assembling prewalls from fair-faced concrete plus a design of a systemic formwork follow. More detailed engineering report of a Spiroll ceiling panels’ assembly is part of the inspection and test plan. Last but not least, an OSH plan for installation works with a mobile crane and a noise study can also be found in this thesis.
9
Reichardt, Christian L. Lange Andrew E. Lange Andrew E. "A high resolution measurement of temperature anisotropies in the cosmic microwave background radiation with the complete ACBAR data set /." Diss., Pasadena, Calif. : California Institute of Technology, 2008. http://resolver.caltech.edu/CaltechETD:etd-12202007-130152.
Full text
10
Oreifej, Omar. "Robust Subspace Estimation Using Low-Rank Optimization. Theory and Applications in Scene Reconstruction, Video Denoising, and Activity Recognition." Doctoral diss., University of Central Florida, 2013. http://digital.library.ucf.edu/cdm/ref/collection/ETD/id/5684.
Full textAbstract:
In this dissertation, we discuss the problem of robust linear subspace estimation using low-rank optimization and propose three formulations of it. We demonstrate how these formulations can be used to solve fundamental computer vision problems, and provide superior performance in terms of accuracy and running time. Consider a set of observations extracted from images (such as pixel gray values, local features, trajectories...etc). If the assumption that these observations are drawn from a liner subspace (or can be linearly approximated) is valid, then the goal is to represent each observation as a linear combination of a compact basis, while maintaining a minimal reconstruction error. One of the earliest, yet most popular, approaches to achieve that is Principal Component Analysis (PCA). However, PCA can only handle Gaussian noise, and thus suffers when the observations are contaminated with gross and sparse outliers. To this end, in this dissertation, we focus on estimating the subspace robustly using low-rank optimization, where the sparse outliers are detected and separated through the `1 norm. The robust estimation has a two-fold advantage: First, the obtained basis better represents the actual subspace because it does not include contributions from the outliers. Second, the detected outliers are often of a specific interest in many applications, as we will show throughout this thesis. We demonstrate four different formulations and applications for low-rank optimization. First, we consider the problem of reconstructing an underwater sequence by removing the turbulence caused by the water waves. The main drawback of most previous attempts to tackle this problem is that they heavily depend on modelling the waves, which in fact is ill-posed since the actual behavior of the waves along with the imaging process are complicated and include several noise components; therefore, their results are not satisfactory. In contrast, we propose a novel approach which outperforms the state-of-the-art. The intuition behind our method is that in a sequence where the water is static, the frames would be linearly correlated. Therefore, in the presence of water waves, we may consider the frames as noisy observations drawn from a the subspace of linearly correlated frames. However, the noise introduced by the water waves is not sparse, and thus cannot directly be detected using low-rank optimization. Therefore, we propose a data-driven two-stage approach, where the first stage “sparsifies” the noise, and the second stage detects it. The first stage leverages the temporal mean of the sequence to overcome the structured turbulence of the waves through an iterative registration algorithm. The result of the first stage is a high quality mean and a better structured sequence; however, the sequence still contains unstructured sparse noise. Thus, we employ a second stage at which we extract the sparse errors from the sequence through rank minimization. Our method converges faster, and drastically outperforms state of the art on all testing sequences. Secondly, we consider a closely related situation where an independently moving object is also present in the turbulent video. More precisely, we consider video sequences acquired in a desert battlefields, where atmospheric turbulence is typically present, in addition to independently moving targets. Typical approaches for turbulence mitigation follow averaging or de-warping techniques. Although these methods can reduce the turbulence, they distort the independently moving objects which can often be of great interest. Therefore, we address the problem of simultaneous turbulence mitigation and moving object detection. We propose a novel three-term low-rank matrix decomposition approach in which we decompose the turbulence sequence into three components: the background, the turbulence, and the object. We simplify this extremely difficult problem into a minimization of nuclear norm, Frobenius norm, and L1 norm. Our method is based on two observations: First, the turbulence causes dense and Gaussian noise, and therefore can be captured by Frobenius norm, while the moving objects are sparse and thus can be captured by L1 norm. Second, since the object's motion is linear and intrinsically different than the Gaussian-like turbulence, a Gaussian-based turbulence model can be employed to enforce an additional constraint on the search space of the minimization. We demonstrate the robustness of our approach on challenging sequences which are significantly distorted with atmospheric turbulence and include extremely tiny moving objects. In addition to robustly detecting the subspace of the frames of a sequence, we consider using trajectories as observations in the low-rank optimization framework. In particular, in videos acquired by moving cameras, we track all the pixels in the video and use that to estimate the camera motion subspace. This is particularly useful in activity recognition, which typically requires standard preprocessing steps such as motion compensation, moving object detection, and object tracking. The errors from the motion compensation step propagate to the object detection stage, resulting in miss-detections, which further complicates the tracking stage, resulting in cluttered and incorrect tracks. In contrast, we propose a novel approach which does not follow the standard steps, and accordingly avoids the aforementioned difficulties. Our approach is based on Lagrangian particle trajectories which are a set of dense trajectories obtained by advecting optical flow over time, thus capturing the ensemble motions of a scene. This is done in frames of unaligned video, and no object detection is required. In order to handle the moving camera, we decompose the trajectories into their camera-induced and object-induced components. Having obtained the relevant object motion trajectories, we compute a compact set of chaotic invariant features, which captures the characteristics of the trajectories. Consequently, a SVM is employed to learn and recognize the human actions using the computed motion features. We performed intensive experiments on multiple benchmark datasets, and obtained promising results. Finally, we consider a more challenging problem referred to as complex event recognition, where the activities of interest are complex and unconstrained. This problem typically pose significant challenges because it involves videos of highly variable content, noise, length, frame size ... etc. In this extremely challenging task, high-level features have recently shown a promising direction as in [53, 129], where core low-level events referred to as concepts are annotated and modeled using a portion of the training data, then each event is described using its content of these concepts. However, because of the complex nature of the videos, both the concept models and the corresponding high-level features are significantly noisy. In order to address this problem, we propose a novel low-rank formulation, which combines the precisely annotated videos used to train the concepts, with the rich high-level features. Our approach finds a new representation for each event, which is not only low-rank, but also constrained to adhere to the concept annotation, thus suppressing the noise, and maintaining a consistent occurrence of the concepts in each event. Extensive experiments on large scale real world dataset TRECVID Multimedia Event Detection 2011 and 2012 demonstrate that our approach consistently improves the discriminativity of the high-level features by a significant margin.Ph.D.
Doctorate
Electrical Engineering and Computing
Engineering and Computer Science
Computer Engineering
11
Bredthauer, Andreas. "Tensionless Strings and Supersymmetric Sigma Models : Aspects of the Target Space Geometry." Doctoral thesis, Uppsala : Acta Universitatis Upsaliensis, 2006. http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-7105.
Full text
12
Blume, Walter Steinbach Fritz Pasternack Jonathan. "Brahms in the Meiningen Tradition : his symphonies and Haydn variations according to the markings of Fritz Steinbach, edited by Walter Blume : a complete translation with background and commentary /." Thesis, Connect to this title online; UW restricted, 2004. http://hdl.handle.net/1773/15510.
Full textAbstract:
Thesis (D. Mus. Arts)--University of Washington, 2004.Vita. Translation of: Brahms in the Meiningen Tradition / herausgegeben von Walter Blume. Includes bibliographical references (leaves 129-131).
13
Persson, Jonas. "Strings as Sigma Models and in the Tensionless Limit." Doctoral thesis, Uppsala : Acta Universitatis Upsaliensis, 2007. http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-7783.
Full text
14
RIU, EMMANUELE. "Dante, Pietro Lombardo e le "Sententiae". Un dibattito intorno a natura, peccato originale, grazia, libertà e predestinazione nella "Commedia"." Doctoral thesis, Università degli studi di Genova, 2020. http://hdl.handle.net/11567/1006856.
Full textAbstract:
La mia tesi di Dottorato mira a indagare i legami tra la Commedia di Dante, le Sententiae di Pietro Lombardo (il più celebre manuale di teologia del Basso Medioevo) e i commenti duecenteschi al testo del Magister sententiarum. Il primo capitolo raccoglie tutte le indicazioni che la critica dantesca e i commentari alle opere di Dante hanno offerto riguardo a quali argomenti del poema possano essere utilmente interpretati attraverso le Sententiae. Il secondo e il terzo capitolo prendono in esame alcuni argomenti specifici: da un lato, il tema del peccato originale, della natura umana prima della Caduta di Adamo ed Eva, e le dinamiche della Redenzione dell’uomo mediante l’Incarnazione di Cristo (Purg. xxvii-xxxiii e Par. vii); dall’altro, i complessi rapporti fra grazia e libertà nel poema, la possibilità (o meno) di salvezza per coloro che, per motivi geografici e/o cronologici, si situano fuori dal circolo della Rivelazione, e il dibattutissimo tema della predestinazione divina (Par. xix-xx e xxxii). Il ricorso alle Sententiae e ai relativi commenti permette sia, in alcuni casi, di rilevare possibili influenze dirette dell’opera teologica sul poema dantesca; sia, in altri, di ricostruire il background teoretico relativo alle questioni elencate, così da comprendere e apprezzare meglio la posizione di Dante in merito.
15
Shi-Ru, You, and 游世儒. "Text extraction in complex background." Thesis, 2001. http://ndltd.ncl.edu.tw/handle/75169127492963355346.
Full textAbstract:
碩士元智大學
資訊工程學系
89
Extracting text region from images or video streams can be widely used in various application fields. A new method for text extraction in complex background is proposed in this study. First, a cellular decomposition quantization method is applied to the color document image. Color and texture featured are then integrated to detect homogeneous textured regions. Finally, different rules of text identification are individually applied to different types of textured regions. Experimental results have been conducted to demonstrate that the divide-and-conquer strategy can cope with the variations of complex background. The experimental results are satisfactory.
16
Po-Wen, Cheng, and 鄭博文. "Exracting Text from Complex Background." Thesis, 1996. http://ndltd.ncl.edu.tw/handle/07477940291962294562.
Full textAbstract:
碩士國立臺灣科技大學
電機工程研究所
84
In document analysis, one important task is to extract text from a complex background. In this study, we develop an algorithm for this task. We first analyze the image, consisting of both text and background. Our algorithm is based on the difference in features between text and the extraneous background and involves four step: (1) block segmentation, (2) classification, (3) region formation, and (4) detecting of textual regions. Also, our method works with a 256 gray scale image directly and requires no binarization. The computationally non-intensive block segmentation and classification procedures yield data which is easily compressible, thus also decreasing the storage requirement.
17
SHARMA, RICHA. "HAND GESTURE RECOGNITION IN COMPLEX BACKGROUND." Thesis, 2013. http://dspace.dtu.ac.in:8080/jspui/handle/repository/15658.
Full textAbstract:
The Hand Gesture Recognition is an important component of the computer vision system and is of immense interest due to its application in tele presence, sign language, intelligent video surveillance system, human computer interaction, virtual reality, traffic control etc. Hand Gesture recognition system is a process involving classifying the given gesture of the hand portion. Our thesis gives a technique for the recognition of hand gesture from the 11 different static gestures taken from NUS hand posture dataset. The purpose of this thesis is to study and develop a method for the efficient detection and classification of hand gestures in the complex background. Hand gesture detection in complex background is seen as a challenging task. Skin similarity measure is used to detect the hand in complex background hand gesture image.
The whole of the image is divided into two classes one is hand and other is background. Then Otsu’s threshold algorithm is used for the segmentation of hand gesture portion in gray scale image. Then the morphological image processing techniques are used to eliminate external and internal noise in the segmented image of the hand portion comprising the hand gesture. Gabor filter is used to find features which are then fed into PCA to get the reduced feature matrix. The features of the shape are obtained by the edges so determined by the canny edge detector. These extracted features are applied as input to classifier. Linear SVM classifier discriminates the images based on dissimilarity between two images. Experimental result shows that 94.6% recognition accuracy is achieved by using linear classifier.
18
Wen-Hsien, Chen, and 陳文賢. "Wavelet-Based Face Detection under Complex Background." Thesis, 2003. http://ndltd.ncl.edu.tw/handle/01136870780535409369.
Full textAbstract:
碩士義守大學
電子工程學系
91
There are more and more application in human face recognition and robot vision with the development of technology recently. For this reason, face detection was became an important technology In this paper, we will detect Human Face by wavelet coefficient and characteristics of different wavelet sub-bands. Moreover, we will provide a ” Extrema Method ” for the extraction of wavelet-based feature. It will be the useful information and help the detection of edge. Finally, we will develop a algorithm for face detection using the information of Extrema. To discuss the efficiency of the face detect algorithm in complex background. Furthermore, the correct detection rate is above 95%.
19
Kao, Chen-Yu, and 高振祐. "Seam Carving for Images with Complex Background." Thesis, 2012. http://ndltd.ncl.edu.tw/handle/74511672676105674767.
Full textAbstract:
碩士淡江大學
資訊工程學系碩士班
100
As displays become commonly used and are incorporated into more and more devices, such as mobile phones, multimedia players, there has been an increased focus on image resizing techniques to fill an image to an arbitrary screen size. Traditional methods such as cropping or scaling might introduce undesirable losses in information or distortion in perception. Recently, content-aware image retargeting method proposed by Avidan and Shamir called “Seam Carving” has good results. In particular, seam carving has gained attention as an effective solution, using simple filters to detect and preserve the high-energy areas of an image. But in some cases this algorithm may not provide satisfactory results, especially, for images having complex background or high-contrast colors. In this study, we would like to improve the existing seam carving methods by selecting more appropriate seams to remove so that it can adapt a variety of environment to provide desired results for users.
20
Lai, Yun-Zhi, and 賴允植. "Human Faces Detection in Colored Complex Background Images." Thesis, 2000. http://ndltd.ncl.edu.tw/handle/15203684468893755254.
Full textAbstract:
碩士國立臺灣科技大學
電機工程系
88
The first step of an automatic human faces identification system is the location of face regions within a captured image. Because human faces are complicated and varied, it is difficult to automatically locate them in a complex background. Besides accurately locating human faces, if we want to build a real-time human faces identification system, the execution time of the system should also be considered. Thus, it is a valuable research topic on the automatic location of human faces at a fast and high correction rate. In this thesis, we propose an efficient faces detection system that can locate human faces in colored complex background images. This faces detection system is based on color information and the geometrical relations of facial features. At the first stage, we extract possible eye blocks by means of edge detection. Then, the number of possible eye blocks is reduced by using the color information. At the next stage, we propose some rules for pairing these blocks to generate possible face candidates by the geometrical relations of facial features. Finally, we again employ the color information and other facial features to eliminate false candidates. The experimental results prove that our approach is feasible and very quickly.
21
Zen, Fu-Jie, and 曾富傑. "Wavelet-Based Single Face Detection under Complex Background." Thesis, 2004. http://ndltd.ncl.edu.tw/handle/13553746346265722050.
Full textAbstract:
碩士國立成功大學
電機工程學系碩博士班
92
Human face recognition has attracted many attentions in research in recent years. This technology is very important in many applications such as human-computer interfaces or security access control. Face detection is the leading phase of face recognition. Locating face correctly and accurately is necessary for high recognition-rate face recognition. This thesis deals with how to locate face correctly in gray-level images with complex backgrounds. The proposed method in this thesis includes 2 stages. The stage 1 extracts extrema from wavelet coefficients of images after wavelet decomposition to locate face. When the stage 1 can not find the face, the stage 2 is included to locate face from the wavelet coefficients of images by thresholding. To improve the accuracy of face location, the method finds the eyes from the located face area by integral projection. The experiments of this thesis use the BioID database, the Visionic database and the database shot by ourselves to test our proposed method. The BioID database consists 1521 images with a resolution of 384×286 pixels. These images show the frontal view of a face of one out of 23 different test persons and shot in some different conditions of expression, lighting, and background. The Visionic database consists of 112 pictures shot by 112 different people. The size of each picture is uncertain. There are 60 images in the databases shot by ourselves, the size of image is 512×384. The processing time of the method with stage 2 is longer than that of the method without stage 2, but the difference is small. The processing time for an image with a resolution of 384×286 pixels is one to two seconds. The average detection rate of the proposed method without stage 2 and with stage 2 are 93% and 96% respectively.
22
Wu, Ming Shan, and 吳旻繕. "A Model-Based Complex Background Gesture Recognition System." Thesis, 1996. http://ndltd.ncl.edu.tw/handle/16230337388669134458.
Full text
23
Luo, GuoYu, and 羅國育. "Gesture recognition based on compact hand segmentation under complex background." Thesis, 2010. http://ndltd.ncl.edu.tw/handle/74140739710320288034.
Full textAbstract:
碩士國立高雄應用科技大學
光電與通訊研究所
98
This thesis presents innovative compact hand segmentation for gesture recognition. First, we devise singular value decomposition based image enhancement technique to remove the dark background, left the remaining pixels of skin color. Then the polynomial approximation YCbCr color model is adopted to extract hand. After alignment, a novel adaptive singular value decomposition is presented for lighting compensation. Finally, we propose using quincunx pyramid sampling algorithm to reduce the impact of variation in gestures, and then the self-eigenhand recognizer with genetic algorithms is constructured to select discriminant eigenvector subset for classification. While on one hand, the approach maximizes the differences amidst the hand images of different gestures, it also minimizes the lighting and pose variations of the same gesture. Experimental results on our database and a live sequence show that our method is superior to conventional method without compact hand segmentation against complex scenes. For the 768 images in the tests set, in the inside testing, the classification achieves the AAR of 99.55% and the FAR of 0.0001%. In live testing, the classification achieves the accuracy rate of 91.7% and the error rate of 8.3%. In terms of speed, our system is faster, and our images size is pixels, operating at an average processing time 1 sec per gesture on an AMD64 Athlon CPU 2.0 GHz PC. Keywords: Compact hand segmentation, singular value decomposition based image enhancement, lighting compensation, quincunx pyramid sampling.
24
Tsai, Tsung Ying, and 蔡宗穎. "An Object Segmentation System Design using Color Analysis with Complex Background." Thesis, 2011. http://ndltd.ncl.edu.tw/handle/88978418398464445042.
Full textAbstract:
碩士東海大學
資訊工程學系
99
This paper presents a block-based image segmentation method. The homogeneity of all the pixels in a block are calculated, and the color information of the block can be obtained. Thus, the background model can be built, and a moving objects can be detected accordingly in a complex background. In an image sequence, some background images are used to capture blocks’ color information to create the background model. Their texture information is applied to process a complex background. The background model can be updated while background changes. The traditional codebook background method useds complex algorithms, and requires more time and memory. Compared to other segmentation methods, our approach with the background model can segment a moving object rapidly and successfully in a complex background environment.
25
Mao, Gang-Zeng, and 毛綱增. "Research on Real-Time Hand Detection and Tracking against Complex Background." Thesis, 2009. http://ndltd.ncl.edu.tw/handle/47810870772013516231.
Full textAbstract:
碩士國立臺灣科技大學
資訊工程系
97
Most hand detection and tracking algorithms can only be applied on fairly simple and similar backgrounds. We propose to combine a modified object detection method proposed by Viola and Jones [4] with the skin-color detection method to perform hand detection and tracking against complex backgrounds. Our experimental results show that the proposed method is effective in near real-time speed (15 frames per second.)
26
Tsai, Jen-Yu, and 蔡任右. "Design and Implementation of Visual-Based Tracking ControlSystems in Complex Background." Thesis, 2009. http://ndltd.ncl.edu.tw/handle/20062656381712956928.
Full textAbstract:
碩士國立成功大學
工程科學系碩博士班
97
The objective of this thesis is to develop a tracking control system based on visual measurement in complex background. A visual system is used to measure the position of the target object and then the motor is used to track the target object. The system consists of an image processing module, image sensor, and tracking controller. In this system, the image processing module is implemented on an FPGA device. The image sensor is a CMOS camera. Besides, the tracking controller is implemented on a digital signal processor. The connected component labeling and size filter methods are used to detect the moving object. However, these methods easily cause errors due to the noise of the image. In order to overcome this difficulty, the median filter and the morphological filter are used to attenuate the effect of the image noise. The system is developed and proven to function well through simulation and experiments.
27
Liu, Yi-chi, and 劉益池. "Implementing High Performance Face Detection Algorithm of Complex Background in SOPC." Thesis, 2006. http://ndltd.ncl.edu.tw/handle/c5zcb9.
Full textAbstract:
碩士國立臺灣科技大學
電子工程系
94
Human face detection plays an important role in many applications such as video surveillance, face recognition, and face image database management. This research utilized the highly integrated SOPC to combine the color face detection so as to implement a high performance face detection algorithm. Altera’s FPGA-based SOPC used in the thesis includes a soft-core 32-bit RISC micro-processor, user-defined hardware modules can be easily combined with the micro-processor. SOPC has advantages in several ways: 1) complete hardware module library, 2) friendly user interface for writing C code, 3) fast hardware execution speed, 4) software can be easily developed, 5) more flexible than ASIC for development and 6) lower power consumption than personal computers. The human face detection and tracking sub-system identifies the possible face region using skin color information， achieving the preliminary positioning of the face. After the preliminary face positioning we locate the face outline using an rectangular mask, and finally detect the eyes and the lip to confirm if the object of interest is indeed the human face. Our method is effective on facial variations such as close eyes, open moth, wear eyeglasses and. a half profile face. The experimental results show that the detection rate is 98.4% in a simple background and 90.8% in a complex one. The processing speed in NIOS testing is 0.418 secs for a frame. Thereby, our proposed algorithm is suitable for using in real-time systems with high performance.
28
Yao, Chung-hao, and 姚中浩. "Embedded real-time system for license plate recognition under complex background." Thesis, 2012. http://ndltd.ncl.edu.tw/handle/81305683171827371536.
Full textAbstract:
碩士國立雲林科技大學
電機工程系碩士班
100
Along with economic development, increased Taiwan''s traffic management needs, such as car park management, as well as schools, building junction control, we need the license plate recognition system to monitor and control out of the vehicle, compared to the traditional license plate recognition system, embedded license plate recognition system at a lower cost and easy to install and other advantages. In this paper, ADSP-BF561 combined with vehicular camera image of the school gate and surveillance camera images on an ordinary road, the development of embedded real-time license plate recognition system. The development plate-form used for the dual-core architecture is divided into A, B, core, so the task allocation, the A core is responsible for license plate location; the B core is responsible for the license plate cutting, and character recognition. In a complex background, we first make the picture mask edge detection, because the license plate edge will show the changes of the vertical and horizontal, in accordance with the license plate texture change frequently, to filter out the license plate of the candidate from the image, then the license plate appearance characteristics as well as three midline crossover rules, filter out the license plate; After license plate location, the B core take trimming, cutting character, the character recognition ; The sharpening of license plate binarization image, plate boundary will be strong numerical changes in boundary features, and so can be based on cutting edge up and down, left and right trimming, retain the character part of the outer contours depict France, cutting characters and regularization of the characters after cutting. Part of Character recognition, according to different classification of the characters in upper and lower contours, and then compared with the contours and texture of the standard characters, identification results. The experimental results show that the license plate location system for the complex background of a success rate of 91.3 percent, trimming success rate of 75.5% character recognition rate of 94.2%, 52.9% of the rear license plate recognition rate; and the system is able to complete a single image within 16ms license plate location and recognition, it can achieve the purpose of real-time processing.
29
Chuang, Wan Li, and 莊萬里. "The Observers’ Performance of Detecting Semantic Images from a Complex Background." Thesis, 2012. http://ndltd.ncl.edu.tw/handle/06758264904082046974.
Full textAbstract:
碩士中國文化大學
資訊傳播學系
101
Due to the drastically growth of visual information in recent decades, people are frequently exposed to vast amount of images. When facing numerous images, identifying any specific image from many irreverent images becomes challenging to the viewer’s visual cognitive capacity. To investigate the behavior of detecting the specific image from complex background, this research mimics a situation happens in everyday life, i.e., to search an image that match a specific theme from the background composed of various images. The researcher carried out two visual searching experiments to examine the effect of certain factors. In the experiment I, the factors including the given image theme, color, visual angle of stimuli size and background distractors were examined. After comparing the correction rate of image theme detection and the self-reported confidence scores, the results show that the different given theme (human action, expressions, arbitrary objects) and viewing angle are two effective factors in correction rate and confidence score, and the other factors are found effective in one of two. The experiment II is also designed as a theme detection task, but theme of target image are constrained in four kinds of facial expressions, happiness, sad, pain and angry. The other factors are similar with that of in experience I. Besides the observation the detection rate and the confidence scores, the detecting behavior of the participants were also recorded by an eye tracker. The results shown that the participants were more sensitive to detect the facial expression representing happiness. The effect of color, viewing angle and distractors types are identical to that of in experiment I. The eye movement pattern, including the fixation time, fixation duration and saccade path, is specifically influenced by the present viewing size of the stimuli.
30
Chia-LungWeng and 翁嘉隆. "Application of Wavelet Transform to Multi-Target Image Analysis in Complex Background." Thesis, 2015. http://ndltd.ncl.edu.tw/handle/91930019255139603806.
Full textAbstract:
碩士國立成功大學
航空太空工程學系碩士在職專班
103
This paper proposes a set of methods for detecting and recognizing multiple target objectives in a complicated image based on wavelet transform. The constructed image recognition system consists of eight steps: a. color model transform, b. wavelet transform, c. noise filter, d. morphological image processing, e. edge detection, f. feature extraction and representation, g. target framing, h. computation of target object center. This study, a full color camera CCD is adopted to capture images, to accurately represent color before processing and separate the complicated background via color detection and locate target objects within certain color range. During wavelet transform, it uses high- and low-pass filtering to reduce the image order, making the height and width half as much as the original size, by which the image process becomes quicker. From the experimental results on some environmental images, the proposed method is able to detect the multiple target objects in complex images quickly and precisely, as well as effectively calculate the pattern information of objects.
31
Shi, Dong-Her, and 施東和. "An Efficient Region-Based Algorithm for Detecting Human Faces in a Complex Background." Thesis, 1996. http://ndltd.ncl.edu.tw/handle/35903699970841679425.
Full textAbstract:
碩士逢甲大學
資訊工程研究所
84
Detecting human faces automatically in a nature scene is a difficult yet significant problem. Due to the complexity and unstability of the pattern of a human face, both accuracy and speed are important factors for detecting human faces in a complex background in a fully automatic human face recognition system. In this paper an efficient region-based algorithm to locate human faces in a complex background in which both the location and the size of a face are unknown is proposed. A hierarchical knowledge-based approachusing region features formed by edge features of the human face is utilizedin our algorithm which is consist of three levels. At level 1, after edgedetection process the local maximum rectangular blanks in an unknown image,regarded as the necessary portions of the possible human faces are found.This approach avoid the full scan of the image with all possible sizes andlocations of human faces. At level 2, the candidate human faces are selectedusing the weighting mask as the global features of a face obtained fromthe machine learning using a lot of human face samples. At level 3, a facefeature vector representing the local features of the eyes, nose, mouth and eyebrows is used for the fine matching with each candidate human face.Experimental results show that the proposed algorithm for detecting andlocating human faces in a complex background is fast and robust. Face detection Face recognition Edge detection Pattern recognition
32
Tan, Jheng-Jhong, and 譚正中. "Missing Objects or Abandoned Objects Detection Method Under Low Brightness and Complex Background." Thesis, 2010. http://ndltd.ncl.edu.tw/handle/3sm9az.
Full textAbstract:
碩士國立臺北科技大學
電子電腦與通訊產業研發碩士專班
98
Moving object detection, stolen valuables finding, and abandoned objects searching, are current trends of research for digital surveillance systems. However, many problems are still not solved. One issue is misjudgment generated in low light environment, and another is how to detect moving objects under complex background. In order to adapt to any changes in the environment, we have to build background model. This study uses Gaussian Mixture Model to build the background model. At first, it has to adapt to low light and effectively detect missing objects and abandoned objects in complex background environments. Noises detected are filtered through morphology filter. Finally, contour detection algorithm is used to determine missing objects or abandoned objects. This kind of digital surveillance systems can be applied to security systems for valuables in department stores, company, defense of military places and home security and thus become an essential part of valuables security and safety considerations.
33
Wang, Chun-Hung, and 王竣弘. "Design and Chip Implementation of Three Dimensional Gestures Interface Based on Complex Background." Thesis, 2014. http://ndltd.ncl.edu.tw/handle/hgf796.
Full textAbstract:
碩士國立臺北科技大學
電腦與通訊研究所
102
In this thesis, we proposed a “Design and Chip Implementation of Three Dimensional Gestures Interface Based on Complex Background.” Our system could be used in illumination changes, varying color and cluttered background. Avoid affecting the system accuracy from the different environment, The system track user’s hand and fingers. Furthermore, the system recognized the state of user’s hand and fingers via information. Our system provided a human-computer interaction(HCI) by using gesture. Our proposed approach is that when a user control the system, the system would detect the eight kinds of moving direction while user adopts 5 fingers gesture. The system would detect the finger of scaling image while user uses 2 fingers gesture. In the scale mode, the system would zoom in or zoom out the image. In addition the system would detect the finger of clicking image while user adopts 1 finger gesture. In finger clicking state, user could select the operation of virtual button by clicking that button. Finally, we design digital chip using hardware architecture and cell-based design to implement our system. When we get the image from camera, our chip would detect the ROI size, center coordinates of hand for next image tracking.
34
Jhang-YiJian and 簡彰億. "Development of a DSP-Based Vision-Guided Omnidirectional Mobile Robot in Complex Background." Thesis, 2010. http://ndltd.ncl.edu.tw/handle/73122008963567207546.
Full textAbstract:
碩士國立成功大學
工程科學系碩博士班
98
The objective of this thesis is to design and realize a visual servo system with complex background. This system is then used to guide an omnidirectional mobile robot to achieve the object tracking. The system uses two image sensors to sense the position of the target in the space, and guides the omnidirectional mobile robot to move to the target. The overall system consists of an omnidirectional mobile robot, image processing modules, two image sensors and tracking controller. The CMOS image sensors are used to acquire the image data. The image processing module and tracking controller are implemented on two different digital signal processors. The connected component labeling method and the size filter are used to differentiate the target object from the complex background, but these two methods are very sensitive to the light changing. In order to solve this problem, we use the morphological methods to improve the ability against the light effects. The state feedback linearization with the PID controller is used to control the vision-guided omnidirectional mobile robot. The system is developed and demonstrated its effectiveness through simulation and experimental results.
35
Tsai, Shiang-Ren, and 蔡翔任. "Efficient Line Detection for Images with Complex Background Using Line-Segment-Priority Approach." Thesis, 2013. http://ndltd.ncl.edu.tw/handle/11007984702190311508.
Full textAbstract:
碩士國立臺灣科技大學
資訊工程系
101
Detecting straight lines from images with complex backgrounds is a real and important problem in pattern recognition and image processing community. However, most of existing line detection algorithms suffer from the accuracy degradation when dealing with images with complex backgrounds. In this paper, a robust and fast line detection algorithm is presented for tackling this kind of images. To suppress the influence of invalid edge points or noise in line detection, we first extract line segments by linking the connected edge points, which will be used as seeds in line detection. For reflecting the importance level of each line segment, we assign an appropriate priority to each line segment. Each time the line segment with the highest priority is selected as a possible line segment which may be promoted to the candidate line segment using the proposed gradient-based voting strategy associated with the reduced Hough space. Finally, a line drawing-based process is applied to check whether the candidate line segment is a true line or not. Based on some real test images with complex backgrounds, experimental results demonstrated the computation-saving and robust advantages of our proposed algorithm for line detection when compared with the previous algorithms, such as the standard Hough transform, the orientation-based elimination algorithm, and the surround suppression-based algorithm.
36
張佑傑. "The Real-Time Face Detection System Under Complex Background and Varying Lighting Condition." Thesis, 2006. http://ndltd.ncl.edu.tw/handle/53806853051876162379.
Full textAbstract:
碩士國立臺灣師範大學
資訊教育學系
94
Human face detection is an important capabilities in a wide range of applications, such as face recognition, face tracking, and content-based image retrieve. Detecting and locating face in image is a necessary procedure before any future processing. We proposed a real-time face detection system including two gradient-based models. In first stage, two Gaussian mixtures of facial and non-facial weighted gradient distribution are used to roughly locate face in image. For accelerating detecting speed, dynamic interval detection algorithm is proposed to avoid redundant computations. In second stage, spatial gradient relation model is proposed to remove false detection and locate the facial positions precisely. In experimental results, weighted gradient distribution and spatial gradient relation model are proven to robust to different facial pose, expression, and rotation. Proposed methods can achieved detection rate of 91% and 95% respectively in database of BioID and Viplab under complex background and varying light condition. Proposed system can detect faces in 10 frames per second with size of 320×240 on a Intel Pentium M 1.5GHz notebook.
37
Lin, Bo-Da, and 林柏達. "A Spatio-Temporal Approach for Video Text Detection and Tracking in Complex Background." Thesis, 2007. http://ndltd.ncl.edu.tw/handle/80541767195959514038.
Full textAbstract:
碩士輔仁大學
電子工程學系
95
Video text detection in unconstrained environment is a great challenge due to arbitrary color, size, orientation and low contrast of text and background. This paper proposes a novel method that not only utilizes texture analysis in spatial domain, but also incorporates temporal information in time domain, to tackle the challenging problem. A 3D wavelet transform is first proposed to filter out high and low frequencies in both spatial and temporal domain. Statistical feature of text are extracted from the filter sub-bands. A Gaussian mixture Bayesian network is derived to classify text regions through the statistical features. Text tracking is achieved by a modified particle filter, which tracks text feature of kernel edge orientation histogram. Experiments are conducted on various real-world videos containing challenging scene text. Performance is compared with existing method. Experimental results are very encouraging and show that the method can detect arbitrary video text in complex background with high accuracy.
38
Ming-De, Lin, and 林明德. "A novel approach for fast detection of specific faces from color images under complex background." Thesis, 2004. http://ndltd.ncl.edu.tw/handle/35479316776386567368.
Full textAbstract:
碩士國立臺灣科技大學
機械工程系
92
This paper proposes a novel fast approach to the human face detection and specific human face recognition in color images under complex background. Human skin information and geometric features are used to select face candidates, then a three-layer backpropagation neural network is used to classify faces and non-faces from face candidates. Experimental results show that the face detetion algorithm can fast detect human faces under various image qualities, complex backgrounds, various lighting conditions, various facial expressions, slightly occluded, one-sided and tilt facial orientations. The eye features are next used to re-locate the precise face image. The expressions, sizes and orientations of the face images that are used to train the neural network for recognize the specific face are not restricted. Lastly, a second three-layer backpropagation neural network is then used to detect if the face is the specific human face. The experiments show that the trained neural networks can efficiently recognize the specific face from images of complex background with satisfactory reliabilities.
39
Devarapalli, Avinash Babu. "Static hand gesture segmentation for images with complex background; detection and tracking of dynamic hand gesture." Thesis, 2013. http://ethesis.nitrkl.ac.in/5204/1/211EC4091.pdf.
Full textAbstract:
This thesis presents color hand gesture segmentation for static images with complex background along with tracking and detection of hand gesture from video sequence. This thesis consists of two works: 1) Static. 2) Dynamic. In the first part, aim is to automatically segment the hand gesture from a given image under different luminance conditions and complex backgrounds. The luminance value affects the color component of an image which leads to increase the noise level in the segmented image. This paper proposes a combined model of two color spaces i.e., HSI, YCbCr and morphological operations with labeling to improve the segmentation performance of color hand gesture from complex backgrounds in terms of completeness and correctness. The proposed color model separates the chrominance and luminance components of the image. The performance of the proposed method is demonstrated through simulation and the experime ntal results reveal that proposed method provides better performance accuracy compared to the HSI and YCbCr methods individually in terms of correctness and completeness. In the second part, aim is to automatic detection and tracking of hand gesture from v ideo sequence under different backgrounds. It involves three steps: 1). Hand tracking 2). Hand detection 3). Hand identification.Here all the simulations are done in MATLAB 10 environment.
40
Chen, Chih-Ping, and 陳治平. "A Study on the Target Separation from Complex Backgrounds." Thesis, 1997. http://ndltd.ncl.edu.tw/handle/85930054767764003966.
Full textAbstract:
碩士國立交通大學
控制工程系
85
In the real-world environment people face, perception is usually complicated andfickle, and this is always worrisome to researchers studying image processing.According to the studies on the field of target separation, most of them considered only simple noise disturbance problems. Recently, many researchers began to studythis question with complicate backgrounds. But the results obtained by these researchersdid not clearly show objects with tiny or unobvious contours. For this reason, we analyze the interaction between backgrounds and target objects. Furthermore, we usemodified edge detection operators to get the target signals. We find that targetsignals may be retained easily after blurring. Therefore, we use this characteristicto obtain target from complex backgrounds. During our experiments, we apply optimal parameter selectiontechnology to reduce false alarms. Finally, our experiment results show that patterns processed byusing our system yield very good target-separation result.
41
Hu, Shyuegong, and 胡學恭. "Multiple-Face Detection & Face Recognition for Complex Backgrounds." Thesis, 2010. http://ndltd.ncl.edu.tw/handle/87643401522123211970.
Full textAbstract:
碩士元智大學
資訊管理學系
98
Human face detection and human face recognition are always popular topics in pattern recognition, they have develop for a long time, and its technique has become more and more robust. At the past, some technique could be restricted by hardware or other experimental settings, so they should made some kind of trade off between detection rate, recognition rate and computing time. Now days, with the development of technology and other equipment some time consuming methods become available. We present a human face detection method which is base on haar-like features and integral image and a human face recognition method which divide human face into different feature area, then use machine learning to divide the feature groups and classify them by these feature information.
42
Cho, Po-Chuan, and 卓柏全. "A Study of Object Tracking in Varying Complex Backgrounds." Thesis, 2009. http://ndltd.ncl.edu.tw/handle/ns63bu.
Full textAbstract:
碩士國立臺北科技大學
自動化科技研究所
97
An object tracking algorithm in a varying complex background, swimming pool, has been studied in this thesis. And its accuracy has been verified by video clips. The proposed approach consists of two main stages: swimmer detection and swimmer tracking. The detection stage begins with employing the mean-shift algorithm to cluster input image, and then chooses the sets by graphical models to train the Gaussian mixture model. Finally, swimmers can be detected in a model-based way. The main purpose in this tracking stage is to compare the accuracy among Kalman filter, mean-shift tracking algorithm, and particle filter algorithms. The experimental results show that the proposed approach can track swimmers efficiently. In the future, we hope to apply it on drowning alarm systems in swimming pools.
43
WANG, SHIH WEI, and 王士維. "A Novel Design of the Face Detection System Under Complex Backgrounds." Thesis, 2011. http://ndltd.ncl.edu.tw/handle/79400747079666122367.
Full textAbstract:
碩士明新科技大學
電子工程研究所
100
The purpose of this research is to present a novel approach for processing face image detection under the complex backgrounds, which usually cause the error detection for face Images. There are objects within background area with the similar color as the face skin or with the eclipse shape as the face contour. The method proposed is to separate the face from the complex backgrounds on the tested images. Firstly the chromatic color messages are normalized and transformed into the color space. In order not to degrade the Influence due to illumination radiation, skin-color segmentation with appropriate thresholding and image binarization are used to separate skin areas from the complex backgrounds to locate the target area as exactly as possible. To delete the noise regions and reconnect the non-continuous ones, dilation and erosion operations of morphological method are utilized to recover the face contour and remove the face noise. Furthermore by the use of face feature, the target face area can be exactly located and at the same time the objects which aren’t the human face can be filtered out completely. The method proposed is demonstrated by the experiments and its performance is achieved as expected.
44
PLACATKOVÁ, Petra. "Subjektivní hodnocení celkového zdravotního stavu pacienta na konci redukční fáze léčby lymfedému." Master's thesis, 2007. http://www.nusl.cz/ntk/nusl-46192.
Full textAbstract:
The graduation thesis is focused on Lymphedema patientś problems. It addresses how the patient values his or her own health state. The differences between a patientś health state before therapy and after finishing its reduction phase have been compared. The following aspect have been valued: limited movement, pain and mental stress. All of them were improved after the reduction phase. Thus Hypothesis 1 has been confirmed: "At the end of the Reduction Phase of Lymphedema therapy the overal health state of the patient is improved." Out of above meditioned three aspect, the mental stress improved at a lowest rate. Differences between patients with and without good social background were demonstrated. Patients with bad social background valued mental stress caused by Lymphedema before and after the therapy as significantly higher then patients with good social background. Thus Hypothesis 2 was also confirmed: "Patients with good social background cope with their disease in a better way." In order to test the hypotheses the quantitative research was done. The technique was standardised integration.
45
Yang, Kai 1982. "FFT and multigrid accelerated integral equation solvers for multi-scale electromagnetic analysis in complex backgrounds." Thesis, 2014. http://hdl.handle.net/2152/26036.
Full textAbstract:
Novel integral-equation methods for efficiently solving electromagnetic problems that involve more than a single length scale of interest in complex backgrounds are presented. Such multi-scale electromagnetic problems arise because of the interplay of two distinct factors: the structure under study and the background medium. Both can contain material properties (wavelengths/skin depths) and geometrical features at different length scales, which gives rise to four types of multi-scale problems: (1) twoscale, (2) multi-scale structure, (3) multi-scale background, and (4) multi-scale-squared problems, where a single-scale structure resides in a different single-scale background, a multi-scale structure resides in a single-scale background, a single-scale structure resides in a multi-scale background, and a multi-scale structure resides in a multi-scale background, respectively. Electromagnetic problems can be further categorized in terms of the relative values of the length scales that characterize the structure and the background medium as (a) high-frequency, (b) low-frequency, and (c) mixed-frequency problems, where the wavelengths/skin depths in the background medium, the structure’s geometrical features or internal wavelengths/skin depths, and a combination of these three factors dictate the field variations on/in the structure, respectively. This dissertation presents several problems arising from geophysical exploration and microwave chemistry that demonstrate the different types of multi-scale problems encountered in electromagnetic analysis and the computational challenges they pose. It also presents novel frequency-domain integral-equation methods with proper Green function kernels for solving these multi-scale problems. These methods avoid meshing the background medium and finding fields in an extended computational domain outside the structure, thereby resolving important complications encountered in type 3 and 4 multi-scale problems that limit alternative methods. Nevertheless, they have been of limited practical use because of their high computational costs and because most of the existing ‘fast integral-equation algorithms’ are not applicable to complex Green function kernels. This dissertation introduces novel FFT, multigrid, and FFT-truncated multigrid algorithms that reduce the computational costs of frequency-domain integral-equation methods for complex backgrounds and enable the solution of unprecedented type 3 and 4 multi-scale problems. The proposed algorithms are formulated in detail, their computational costs are analyzed theoretically, and their features are demonstrated by solving benchmark and challenging multi-scale problems.text
46
Spiridon, Andrei. "Determinação do background geoquímico em metais pesados de solos basálticos do Complexo Vulcânico de Lisboa." Master's thesis, 2019. http://hdl.handle.net/10362/75824.
Full textAbstract:
Devido à crescente urbanização dos meios naturais, a qualidade dos solos é cada vez mais influenciada pelas atividades antrópicas. Assim, no âmbito da preocupação relacionada com a potencial contaminação de solos, torna-se fundamental identificar o estado natural dos parâmetros geoquímicos desses solos de modo a estabelecer uma referência que permita, futuramente, quantificar a influência antrópica.
O objetivo desta dissertação consiste assim em estabelecer uma metodologia para a determinação de níveis de referência para concentrações de metais pesados existentes nos solos, utilizando a fluorescência de raios-X (FRX), uma técnica de análise química, que conta com uma preparação expedita da amostra. Para a exemplificação da metodologia foi efetuado o estudo de solos basálticos formados sobre o Complexo Vulcânico de Lisboa.
Foram executadas sondagens em 26 locais que possibilitaram a recolha de solos em dois níveis de profundidade: Nível T – “Topo” (0,5 – 1,5 m), Nível B – “Base” (2 -3 m). Foi analisada a composição química da fração granulométrica fina (inferior a 63 μm) e da fração granulométrica total (inferior a 2 mm). A análise química possibilitou obter informação sobre uma extensa gama de elementos, com especial destaque para os metais pesados V, Cr, Ba, Ni, Zn, Cu e Pb. A comparação dos resultados obtidos por FRX com resultados obtidos para alguns destes elementos (V, Cr, Pb e Zn) pelo método de espectrometria de emissão atômica (ICP-AES), mostra uma concordância entre os pares de medições dos elementos V, Cr e Pb; para o par de medições referente ao elemento Zn, observa-se uma ausência de relação entre os resultados obtidos pelos dois métodos.
A variabilidade composicional dos solos basálticos na área de investigação não permite definir um limite claro entre as concentrações geogénicas dos metais pesados e as de influências antropogénica. Para o estabelecimento do background geoquímico foi calculado um limite superior para os elementos V, Cr, Ba, Ni, Zn, Cu e Pb utilizando o percentil 95 das curvas de frequência acumulada das concentrações medidas.
47
Reichardt, Christian L. "A High Resolution Measurement of Temperature Anisotropies in the Cosmic Microwave Background Radiation with the Complete ACBAR Data Set." Thesis, 2008. https://thesis.library.caltech.edu/5090/1/thesis.0.1.0.pdf.
Full textAbstract:
The Arcminute Cosmology Bolometer Array Receiver (ACBAR) is a 16-element spiderweb bolometer array operating at 150 GHz. Mounted on the 2.1m Viper telescope, ACBAR has dedicated four years to cosmic microwave background (CMB) observations at the South Pole. We describe the focal plane reconstruction and performance of ACBAR for the 2005 austral winter. We present a new CMB temperature anisotropy power spectrum for the complete ACBAR data set. The addition of data from the 2005 observing season expands the data set by 210%; and the sky coverage by 490%; over the previous ACBAR releases. The expanded data set allows us to derive a new set of band-power measurements with finer ℓ-resolution and dramatically smaller uncertainties. In particular, the band-power uncertainties have been reduced by more than a factor of two on angular scales encompassing the third and fourth acoustic peaks and the damping tail of the CMB power spectrum. The calibration has been significantly improved from 6%; to 2.2%; in temperature by using a direct comparison between CMB anisotropy maps measured by WMAP3 and ACBAR to transfer the WMAP dipole-based calibration to ACBAR. The resulting power spectrum is consistent with theoretical predications for a spatially flat, dark energy-dominated Lambda CDM cosmology.