Rozprawy doktorskie na temat „Prédiction de vidéos”
Utwórz poprawne odniesienie w stylach APA, MLA, Chicago, Harvard i wielu innych
Sprawdź 18 najlepszych rozpraw doktorskich naukowych na temat „Prédiction de vidéos”.
Przycisk „Dodaj do bibliografii” jest dostępny obok każdej pracy w bibliografii. Użyj go – a my automatycznie utworzymy odniesienie bibliograficzne do wybranej pracy w stylu cytowania, którego potrzebujesz: APA, MLA, Harvard, Chicago, Vancouver itp.
Możesz również pobrać pełny tekst publikacji naukowej w formacie „.pdf” i przeczytać adnotację do pracy online, jeśli odpowiednie parametry są dostępne w metadanych.
Przeglądaj rozprawy doktorskie z różnych dziedzin i twórz odpowiednie bibliografie.
Marat, Sophie. "Modèles de saillance visuelle par fusion d'informations sur la luminance, le mouvement et les visages pour la prédiction de mouvements oculaires lors de l'exploration de vidéos". Phd thesis, Grenoble INPG, 2010. http://tel.archives-ouvertes.fr/tel-00497787.
Pełny tekst źródłaMarat, Sophie. "Modèles de saillance visuelle par fusion d'informations sur la luminance, le mouvement et les visages pour la prédiction de mouvements oculaires lors de l'exploration de vidéos". Phd thesis, Grenoble INPG, 2010. http://www.theses.fr/2010INPG0012.
Pełny tekst źródłaHuang, Bihong. "Second-order prediction and residue vector quantization for video compression". Thesis, Rennes 1, 2015. http://www.theses.fr/2015REN1S026/document.
Pełny tekst źródłaVideo compression has become a mandatory step in a wide range of digital video applications. Since the development of the block-based hybrid coding approach in the H.261/MPEG-2 standard, new coding standard was ratified every ten years and each new standard achieved approximately 50% bit rate reduction compared to its predecessor without sacrificing the picture quality. However, due to the ever-increasing bit rate required for the transmission of HD and Beyond-HD formats within a limited bandwidth, there is always a requirement to develop new video compression technologies which provide higher coding efficiency than the current HEVC video coding standard. In this thesis, we proposed three approaches to improve the intra coding efficiency of the HEVC standard by exploiting the correlation of intra prediction residue. A first approach based on the use of previously decoded residue shows that even though gains are theoretically possible, the extra cost of signaling could negate the benefit of residual prediction. A second approach based on Mode Dependent Vector Quantization (MDVQ) prior to the conventional transformed scalar quantization step provides significant coding gains. We show that this approach is realistic because the dictionaries are independent of QP and of a reasonable size. Finally, a third approach is developed to modify dictionaries gradually to adapt to the intra prediction residue. A substantial gain is provided by the adaptivity, especially when the video content is atypical, without increasing the decoding complexity. In the end we get a compromise of complexity and gain for a submission in standardization
Franceschi, Jean-Yves. "Apprentissage de représentations et modèles génératifs profonds dans les systèmes dynamiques". Electronic Thesis or Diss., Sorbonne université, 2022. http://www.theses.fr/2022SORUS014.
Pełny tekst źródłaThe recent rise of deep learning has been motivated by numerous scientific breakthroughs, particularly regarding representation learning and generative modeling. However, most of these achievements have been obtained on image or text data, whose evolution through time remains challenging for existing methods. Given their importance for autonomous systems to adapt in a constantly evolving environment, these challenges have been actively investigated in a growing body of work. In this thesis, we follow this line of work and study several aspects of temporality and dynamical systems in deep unsupervised representation learning and generative modeling. Firstly, we present a general-purpose deep unsupervised representation learning method for time series tackling scalability and adaptivity issues arising in practical applications. We then further study in a second part representation learning for sequences by focusing on structured and stochastic spatiotemporal data: videos and physical phenomena. We show in this context that performant temporal generative prediction models help to uncover meaningful and disentangled representations, and conversely. We highlight to this end the crucial role of differential equations in the modeling and embedding of these natural sequences within sequential generative models. Finally, we more broadly analyze in a third part a popular class of generative models, generative adversarial networks, under the scope of dynamical systems. We study the evolution of the involved neural networks with respect to their training time by describing it with a differential equation, allowing us to gain a novel understanding of this generative model
Luc, Pauline. "Apprentissage autosupervisé de modèles prédictifs de segmentation à partir de vidéos". Thesis, Université Grenoble Alpes (ComUE), 2019. http://www.theses.fr/2019GREAM024/document.
Pełny tekst źródłaPredictive models of the environment hold promise for allowing the transfer of recent reinforcement learning successes to many real-world contexts, by decreasing the number of interactions needed with the real world.Video prediction has been studied in recent years as a particular case of such predictive models, with broad applications in robotics and navigation systems.While RGB frames are easy to acquire and hold a lot of information, they are extremely challenging to predict, and cannot be directly interpreted by downstream applications.Here we introduce the novel tasks of predicting semantic and instance segmentation of future frames.The abstract feature spaces we consider are better suited for recursive prediction and allow us to develop models which convincingly predict segmentations up to half a second into the future.Predictions are more easily interpretable by downstream algorithms and remain rich, spatially detailed and easy to obtain, relying on state-of-the-art segmentation methods.We first focus on the task of semantic segmentation, for which we propose a discriminative approach based on adversarial training.Then, we introduce the novel task of predicting future semantic segmentation, and develop an autoregressive convolutional neural network to address it.Finally, we extend our method to the more challenging problem of predicting future instance segmentation, which additionally segments out individual objects.To deal with a varying number of output labels per image, we develop a predictive model in the space of high-level convolutional image features of the Mask R-CNN instance segmentation model.We are able to produce visually pleasing segmentations at a high resolution for complex scenes involving a large number of instances, and with convincing accuracy up to half a second ahead
Poulin, Roxanne. "Prédiction de l'expérience plaisante en fonction de la performance, la difficulté et la familiarité au jeu en contexte de jeux vidéo". Master's thesis, Université Laval, 2018. http://hdl.handle.net/20.500.11794/30224.
Pełny tekst źródłaGoncalves, Gomes Danielo. "Un modèle connexionniste pour la prédiction et l'optimisation de la bande passante : Approche basée sur la nature autosimilaire du trafic vidéo". Evry-Val d'Essonne, 2004. http://www.theses.fr/2004EVRY0021.
Pełny tekst źródłaThe objective of this thesis is the bandwidth forecasting optimization of a MPEG-4 video flow aggregate in a scenario of provisioning of Video on Demand (VoD) service over Internet. The proposed approach takes into account the self-similar nature of the IP traffic to estimate the Hurst parameter. This metric characterizes the degree of self-similarity of a process such as Internet traffic. A first contribution of this thesis is the design and implementation of a connectionist model which estimates and predicts the Husrt parameter of an aggregate video traffic. A new model called Predictive Connectionist Model (PCM) has been defined and is trained with MPEG traces patterns. The estimation of the bandwidth utilisation is achieved using two prediction techniques which evaluated and compared. Another contribution of this thesis is the integration of the Predictive Connectionist Model in a dynamic provisioning system between a VoD provider and an ISP (Internet Service Provider). This system is designed according to the Policy Based Management architecture of IETF (Internet Engineering Task Force) and is implemented using Web technologies
Favreau, Laurent. "Modélisation du mouvement des quadrupèdes à partir de la vidéo". Phd thesis, Grenoble INPG, 2006. http://tel.archives-ouvertes.fr/tel-00379321.
Pełny tekst źródłaThiesse, Jean-Marc. "Codage vidéo flexible par association d'un décodeur intelligent et d'un encodeur basé optimisation débit-distorsion". Nice, 2012. http://www.theses.fr/2012NICE4058.
Pełny tekst źródłaThis Ph. D. Thesis deals with the improvement of video compression efficiency. Both conventional and breakthrough approaches are investigated in order to propose efficient methods for Intra and Inter coding dedicated to next generations video coding standards. Two tools are studied for the conventional approach. First, syntax elements are cleverly transmitted using a data hiding based method which allows embedding indices into the luminance and chrominance residuals in an optimal way, rate-distortion wise. Secondly, the large motion redundancies are exploited to improve the motion vectors coding. After a statistical analysis of the previously used vectors, an accurate forecast is performed to favor some vector residuals during a last step which modifies the original residual distribution. 90% of the coded vectors are efficiently forecasted by this method which helps to significantly reduce their coding cost. The breakthrough approach comes from the observation of the H. 264/AVC standard and its successor HEVC which are based on a predictive scheme with multiple coding choices, consequently future improvements shall improve texture by extensively using the competition between many coding modes. However, such schemes are bounded by the cost generated by the signaling flags and therefore it is required to transfer some decisions to the decoder side. A framework based on the determination of encoding parameters at both encoder and decoder side is consequently proposed and applied to Intra prediction modes on the one hand, and to the emerging theory of compressed sensing on the other hand. Promising results are reported and confirm the potential of such an innovative solution
Moinard, Matthieu. "Codage vidéo hybride basé contenu par analyse/synthèse de données". Phd thesis, Telecom ParisTech, 2011. http://tel.archives-ouvertes.fr/tel-00830924.
Pełny tekst źródłaChanguel, Nesrine. "Régulation de la qualité lors de la transmission de contenus vidéo sur des canaux sans fils". Phd thesis, Université Paris Sud - Paris XI, 2011. http://tel.archives-ouvertes.fr/tel-00659806.
Pełny tekst źródłaAlaoui, Fdili Othmane. "Optimisation multicritères de la qualité de service dans les réseaux de capteurs multimédia sans fil". Thesis, Valenciennes, 2015. http://www.theses.fr/2015VALE0016/document.
Pełny tekst źródłaThanks to the valuable advances in Micro Electro-Mechanical Systems coupled with their convergence to wireless communication systems, the Wireless Sensor Networks (WSN). In the WSN context, all the efforts are made in order to propose energy-efficient solutions. With the recent developments in CMOS technology, low-cost imaging sensors have been developed. As a result, a new derivative of the WSN, which is the Wireless Video Sensor Network (WVSN), has been proposed. The particularities of the video data as well as the inherent constraints of the nodes have introduced new challenges. In this thesis, we propose two cross-layer based solutions for video delivery over the WVSN. The first solution proposes a new energy efficient and adaptive video compression scheme dedicated to the WVSNs, based on the H.264/AVC video compression standard. The video stream is then handled by an enhanced version of MMSPEED protocol, that we propose and note EQBSA-MMSPEED. Performance evaluation shows that the lifetime of the network is extended by 33%, while improving the video quality of the received stream by 12%. In the second solution, we enrich our compression scheme with mathematical models to predict the energy consumption and the video distortion during the encoding and the transmission phases. The video stream is then handled by a novel energy efficient and improved reliability routing protocol, that we note ERMM. Compared to a basic approach, this solution is extending the network lifetime by 15%, while improving the quality of the received video stream by 35%
Gouta, Ali. "Caching and prefetching for efficient video services in mobile networks". Thesis, Rennes 1, 2015. http://www.theses.fr/2015REN1S001/document.
Pełny tekst źródłaRecently, cellular networks have witnessed a phenomenal growth of traffic fueled by new high speed broadband cellular access technologies. This growth is in large part driven by the emergence of the HTTP Adaptive Streaming (HAS) as a new video delivery method. In HAS, several qualities of the same videos are made available in the network so that clients can choose the quality that best fits their bandwidth capacity. This strongly impacts the viewing pattern of the clients, their switching behavior between video qualities, and thus beyond on content delivery systems. In this context, we provide an analysis of a real HAS dataset collected in France and provided by the largest French mobile operator. Firstly, we analyze and model the viewing patterns of VoD and live streaming HAS sessions and we propose a new cache replacement strategy, named WA-LRU. WA-LRU leverages the time locality of video segments within the HAS content. We show that WA-LRU improves the performance of the cache. Second, we analyze and model the adaptation logic between the video qualities based on empirical observations. We show that high switching behaviors lead to sub optimal caching performance, since several versions of the same content compete to be cached. In this context we investigate the benefits of a Cache Friendly HAS system (CF-DASH) which aims at improving the caching efficiency in mobile networks and to sustain the quality of experience of mobile clients. Third, we investigate the mobile video prefetching opportunities. We show that CPSys can achieve high performance as regards prediction correctness and network utilization efficiency. We further show that CPSys outperforms other prefetching schemes from the state of the art. At the end, we provide a proof-of-concept implementation of our prefetching system
Chaabouni, Souad. "Etude et prédiction d'attention visuelle avec les outils d'apprentissage profond en vue d'évaluation des patients atteints des maladies neuro-dégénératives". Thesis, Bordeaux, 2017. http://www.theses.fr/2017BORD0768/document.
Pełny tekst źródłaThis thesis is motivated by the diagnosis and the evaluation of the dementia diseasesand with the aim of predicting if a new recorded gaze presents a complaint of thesediseases. Nevertheless, large-scale population screening is only possible if robust predictionmodels can be constructed. In this context, we are interested in the design and thedevelopment of automatic prediction models for specific visual content to be used in thepsycho-visual experience involving patients with dementia (PwD). The difficulty of sucha prediction lies in a very small amount of training data.Visual saliency models cannot be founded only on bottom-up features, as suggested byfeature integration theory. The top-down component of human visual attention becomesprevalent as human observers explore the visual scene. Visual saliency can be predictedon the basis of seen data. Deep Convolutional Neural Networks (CNN) have proven tobe a powerful tool for prediction of salient areas in static images. In order to constructan automatic prediction model for the salient areas in natural and intentionally degradedvideos, we have designed a specific CNN architecture. To overcome the lack of learningdata we designed a transfer learning scheme derived from bengio’s method. We measureits performances when predicting salient regions. The obtained results are interestingregarding the reaction of normal control subjects against degraded areas in videos. Thepredicted saliency map of intentionally degraded videos gives an interesting results comparedto gaze fixation density maps and other reference models
Pau, Grégoire. "Ondelettes et décompositions spatio-temporelles avancées : application au codage vidéo scalable". Phd thesis, Télécom ParisTech, 2006. http://pastel.archives-ouvertes.fr/pastel-00002189.
Pełny tekst źródłaBrault, Patrice. "Estimation de mouvement et segmentation d'image". Paris 11, 2005. https://tel.archives-ouvertes.fr/tel-00011310.
Pełny tekst źródłaThe first part of this thesis presents a new vision of the motion estimation (ME) in video sequences. We investigate motion estimation with redundant wavelet families tuned to different kind of transformations and, in particular, to speed. Today video compression standards are supposed to realize the compression in an object-based approach, but still compute raw motion vectors on “blocks”. We thus implemented these wavelet families because 1) they are built to perform motion parameter quantization on several kinds of motions (rotation, speed, acceleration) and 2) based on the motion parameters, we can propose an approach of the ME through the identification of the objects trajectories. The global approach is then closer to a contextual compression, based on the understanding of the scene. The second part introduces two new developments on unsupervised segmentation in a Bayesian approach. 1) we reduce the computation time of a sequence through an iterative implementation of the segmentation. We show an application with the ME of a segmented region. 2) We reduce the segmentation time by making the projection of the image in the wavelet domain. These two developments are based on a Potts-Markov modelling (PMRF) for the labels of the pixels and of the wavelet coefficients. They use a Markov Chain Monte Carlo iterative algorithm with a Gibbs sampler. We also develop a Potts model in the wavelet domain to tune it to the specific orientations of the wavelet subbands
Abdallah, Raed. "Intelligent crime detection and behavioral pattern mining : a comprehensive study". Electronic Thesis or Diss., Université Paris Cité, 2023. http://www.theses.fr/2023UNIP7031.
Pełny tekst źródłaIn the face of a rapidly evolving criminal landscape, law enforcement agencies (LEAs) grapple with escalating challenges in contemporary criminal investigations. This PhD thesis embarks on a transformative exploration, encouraged by an urgent need to revolutionize investigative methodologies and arm LEAs with state-of-the-art tools to combat crime effectively. Rooted in this imperative motivation, the research meticulously navigates diverse data sources, including the intricate web of social media networks, omnipresent video surveillance systems, and expansive online platforms, recognizing their fundamental roles in modern crime detection. The contextual backdrop of this research is the pressing demand to empower LEAs with advanced capabilities in intelligent crime detection. The surge in digital interactions necessitates a paradigm shift, compelling researchers to delve deep into the labyrinth of social media, surveillance footage, and online data. This context underscores the urgency to fortify law enforcement strategies with cutting-edge technological solutions. Motivated by urgency, the thesis focuses on three core objectives: firstly, automating suspect identification through the integration of data science, big data tools, and ontological models, streamlining investigations and empowering law enforcement with advanced inference rules; secondly, enabling real-time detection of criminal events within digital noise via intricate ontological models and advanced inference rules, providing actionable intelligence and supporting informed decision-making for law enforcement; and thirdly, enhancing video surveillance by integrating advanced deep learning algorithms for swift and precise detection of knife-related crimes, representing a pioneering advancement in video surveillance technology. Navigating this research terrain poses significant challenges. The integration of heterogeneous data demands robust preprocessing techniques, enabling the harmonious fusion of disparate data types. Real-time analysis of social media intricacies necessitates ontological models adept at discerning subtle criminal nuances within the digital tapestry. Moreover, designing Smart Video Surveillance Systems necessitates the fusion of state-of-the-art deep learning algorithms with real-time video processing, ensuring both speed and precision in crime detection. Against these challenges, the thesis contributes innovative solutions at the forefront of contemporary crime detection technology. The research introduces ICAD, an advanced framework automating suspect identification and revolutionizing investigations. CRI-MEDIA tackles social media crime challenges using a streamlined process and enriched criminal ontology. Additionally, SVSS, a Smart Video Surveillance System, swiftly detects knife-related crimes, enhancing public safety. Integrating ICAD, CRI-MEDIA, and SVSS, this work pioneers intelligent crime detection, empowering law enforcement with unprecedented capabilities in the digital age. Critical to the integrity of the research, the proposed methodologies undergo rigorous experimentation in authentic criminal scenarios. Real-world data gathered from actual investigations form the crucible wherein ICAD, CRI-MEDIA, and SVSS are tested. These experiments serve as a litmus test, affirming not only the viability of the proposed solutions but also offering nuanced insights for further refinement. The results underscore the practical applicability of these methodologies, their adaptability in diverse law enforcement contexts, and their role in enhancing public safety and security
Mathonat, Romain. "Rule discovery in labeled sequential data : Application to game analytics". Thesis, Lyon, 2020. http://www.theses.fr/2020LYSEI080.
Pełny tekst źródłaIt is extremely useful to exploit labeled datasets not only to learn models and perform predictive analytics but also to improve our understanding of a domain and its available targeted classes. The subgroup discovery task has been considered for more than two decades. It concerns the discovery of rules covering sets of objects having interesting properties, e.g., they characterize a given target class. Though many subgroup discovery algorithms have been proposed for both transactional and numerical data, discovering rules within labeled sequential data has been much less studied. In that context, exhaustive exploration strategies can not be used for real-life applications and we have to look for heuristic approaches. In this thesis, we propose to apply bandit models and Monte Carlo Tree Search to explore the search space of possible rules using an exploration-exploitation trade-off, on different data types such as sequences of itemset or time series. For a given budget, they find a collection of top-k best rules in the search space w.r.t chosen quality measure. They require a light configuration and are independent from the quality measure used for pattern scoring. To the best of our knowledge, this is the first time that the Monte Carlo Tree Search framework has been exploited in a sequential data mining setting. We have conducted thorough and comprehensive evaluations of our algorithms on several datasets to illustrate their added-value, and we discuss their qualitative and quantitative results. To assess the added-value of one or our algorithms, we propose a use case of game analytics, more precisely Rocket League match analysis. Discovering interesting rules in sequences of actions performed by players and using them in a supervised classification model shows the efficiency and the relevance of our approach in the difficult and realistic context of high dimensional data. It supports the automatic discovery of skills and it can be used to create new game modes, to improve the ranking system, to help e-sport commentators, or to better analyse opponent teams, for example