Dans cette étude, nous utilisons des caractéristiques locales/globales en vue d’améliorer la chaîne de transmission des séquences de vidéos. Ce travail est divisé en quatre parties principales qui mettent à profit les caractéristiques de contenu vidéo. La première partie introduit un modèle de prédiction de paramètres d’un encodeur basé sur la complexité du contenu. Ce modèle utilise le débit, la distorsion, ainsi que la complexité de différentes configurations de paramètres afin d’obtenir des valeurs souhaitables (recommandées) de paramètres d’encodage. Nous identifions ensuite le lien en les caractéristiques du contenu et ces valeurs recommandées afin de construire le modèle de prédiction. La deuxième partie illustre le schéma de l’encodage à description multiple (Multiple Description Coding ou MDC, en anglais) que nous proposons dans ces travaux. Celui-ci est optimisé pour des MDC d’ordre-hauts. Le décodage correspondant et la procédure de récupération de l’erreur contenu-dépendant sont également étudiés et identifiés. La qualité de la vidéo reçue a été évaluée subjectivement. En analysant les résultats des expériences subjectives, nous introduisons alors un schéma adaptatif, c’est-à-dire adapté à la connaissance du contenu vidéo. Enfin, nous avons simulé un scénario d’application afin d’évaluer un taux de débit réaliste. Dans la troisième partie, nous utilisons une carte de déplacement, calculées au travers des propriétés de mouvement du contenu vidéo, comme entrée pour l’algorithme de masquage d’erreur par recouvrement (inpainting based error concealment algorithm). Une expérience subjective a été conduite afin d’évaluer l’algorithme et d’étudier la perturbation de l’observateur au visionnage de la vidéo traitée. La quatrième partie possèdent deux sous-parties. La première se penche sur les algorithmes de sélections par HRC pour les grandes bases de données de vidéos. La deuxième partie introduit l’évaluation de la qualité vidéo utilisant la connaissance du contenu global non-référencé. / In this work, the global/local content characteristics are utilized in order to improve the delivery chain of the video sequences. The work is divided into four main parts that take advantages of video content features. The first part introduces a joint content-complexity encoder parameters prediction model. This model uses bitrate, distortion, and complexity of different parameters configurations in order to get the recommended encoder parameters value. Then, the links between content features and the recommended values are identified. Finally, the prediction model is built using these features and the recommended encoder parameter values. The second part illustrates the proposed multiple description coding (MDC) scheme that is optimized for high-order MDC. The corresponding decoding and content-dependent error recovery procedures are also identified. The quality of the received videos is evaluated subjectively. By analyzing the subjective experiment results, an adaptive, i.e. content-aware, scheme is introduced. Finally, an application scenario is simulated to study the realistic bitrate consumption. The third part uses the motion properties of a content to introduce a motion map that will be used as an input for the modified state-of-the-art inpainting based error concealment algorithm. A subjective experiment was conducted to evaluate the algorithm and also to study the content-aware observer’s disturbance when perceiving the processed videos. The fourth part has two sub-parts, the first one is about HRC selection algorithms for the large-scale video database with an improved performance evaluation measures for video quality assessment algorithms using training and validation sets. The second part introduces global content aware no-reference video quality assessment.

3D video is a technology that has seen a tremendous attention in the recent years. Multiview Video Coding (MVC) is an extension of the popular H.264 video coding standard and is commonly used to compress 3D videos. It offers an improvement of 20% to 50% in compression efficiency over simulcast encoding of multiview videos using the conventional H.264 video coding standard. However, there are two important problems associated with it: (i) its superior compression performance comes at the cost of significantly higher computational complexity which hampers the real-world realization of MVC encoder in applications such as 3D live broadcasting and interactive Free Viewpoint Television (FTV), and (ii) compressed 3D videos can suffer from packet loss during transmission, which can degrade the viewing quality of the 3D video at the decoder. This thesis aims to solve these problems by presenting techniques to reduce the computational complexity of the MVC encoder and by proposing a consistent error concealment technique for frame losses in 3D video transmission. The thesis first analyses the complexity of the MVC encoder. It then proposes two novel techniques to reduce the complexity of motion and disparity estimation. The first method achieves complexity reduction in the disparity estimation process by exploiting the relationship between temporal levels, type of macroblocks and search ranges while the second method achieves it by exploiting the geometrical relation- ship between motion and disparity vectors in stereo frames. These two methods are then combined with other state-of-the-art methods in a unique framework where gains add up. Experimental results show that the proposed low-complexity framework can reduce the encoding time of the standard MVC encoder by over 93% while maintaining similar compression efficiency performance. The addition of new View Synthesis Prediction (VSP) modes to the MVC encoding framework improves the compression efficiency of MVC. However, testing additional modes comes at the cost of increased encoding complexity. In order to reduce the encoding complexity, the thesis, next, proposes a bayesian early mode decision technique for a VSP enhanced MVC coder. It exploits the statistical similarities between the RD costs of the VSP SKIP mode in neighbouring views to terminate the mode decision process early. Results indicate that the proposed technique can reduce the encoding time of the enhanced MVC coder by over 33% at similar compression efficiency levels. Finally, compressed 3D videos are usually required to be broadcast to a large number of users where transmission errors can lead to frame losses which can degrade the video quality at the decoder. A simple reconstruction of the lost frames can lead to inconsistent reconstruction of the 3D scene which may negatively affect the viewing experience of a user. In order to solve this problem, the thesis proposes, at the end, a consistency model for recovering frames lost during transmission. The proposed consistency model is used to evaluate inter-view and temporal consistencies while selecting candidate blocks for concealment. Experimental results show that the proposed technique is able to recover the lost frames with high consistency and better quality than two standard error concealment methods and a baseline technique based on the boundary matching algorithm.

H.264/AVC is the state-of-the-art video coding standard which promises to achieve same video quality at about half the bit rate of previous standards (H.263, MPEG-2). This tremendous achievement in compression and perceptual quality is due to the inclusion of various innovative tools. These tools are highly complex and data intensive as a result poses very heavy computational burden on the processors. De-blocking filter is one among them, it is the most time consuming part of the H.264/AVC reference decoder. In this thesis, a performance analysis of the de-blocking filter is made on Intel Pentium 4 processor and accordingly various optimization techniques have been studied and implemented. For some techniques statistical analysis of video data is done and according to the results obtained optimization is performed and for other techniques SIMD instructions has been used to achieve the optimization. Comparison of optimized techniques using SIMD with the reference software has shown significant speedup thus contributing to the real time implementation of the de-blocking filter on general purpose platform. / De-blocking Filter is the most time consuming part of the H.264 High Profile decoder. The process of De-block filtering specified in the H.264/AVC standard is sequential thus not computationally optimal. In this thesis various optimization algorithms have been studied and implemented. When compared to JM13.2 boundary strength algorithm, Static and ICME algorithms are quite primitive as a result no performance gain is achieved, in fact there is a decrease in performance. This dismal performance is due to various reasons, prominent among them are increased memory access, unrolling of loop to 4x4 boundary and early detection of intra blocks. When it comes to the optimization algorithms of Edge filtering module both the algorithms (SIMD and fast algorithm) showed significant improvement in performance when compared to JM13.2 edge filtering algorithm. This improvement is mainly due to the parallel filtering operation done in edge filtering module. Therefore, by using SSE2 instructions large speed up could be achieved on general purpose processors like Intel, while keeping the conformance with the standard.

by Alan Yeung. / Thesis (M.Phil.)--Chinese University of Hong Kong, 1998. / Includes bibliographical references (leaves 53-[56]). / Abstract also in Chinese. / Chapter 1 --- Introduction --- p.1 / Chapter 2 --- MPEG Overview --- p.5 / Chapter 3 --- Framework of Frame-Skipping Lossy Aggregation --- p.10 / Chapter 3.1 --- Video Frames Delivery using Round-Robin Scheduling --- p.10 / Chapter 3.2 --- Underflow Safety Margin on Receiver Buffers --- p.12 / Chapter 3.3 --- Algorithm in Frame-Skipping Aggregation Controller --- p.13 / Chapter 4 --- Replacement of Skipped Frames in MPEG Sequence --- p.17 / Chapter 5 --- Subjective Assessment Test on Frame-Skipped Video --- p.21 / Chapter 5.1 --- Test Settings and Material --- p.22 / Chapter 5.2 --- Choice of Test Methods --- p.23 / Chapter 5.3 --- Test Procedures --- p.25 / Chapter 5.4 --- Test Results --- p.26 / Chapter 6 --- Performance Study --- p.29 / Chapter 6.1 --- Experiment 1: Number of Supportable Streams --- p.31 / Chapter 6.2 --- Experiment 2: Frame-Skipping Rate When Multiplexing on a Leased T3 Link --- p.33 / Chapter 6.3 --- Experiment 3: Bandwidth Usage --- p.35 / Chapter 6.4 --- Experiment 4: Optimal USMT --- p.38 / Chapter 7 --- Implementation Considerations --- p.41 / Chapter 8 --- Conclusions --- p.45 / Chapter A --- The Construction of Stuffed Artificial B Frame --- p.48 / Bibliography --- p.53

The growing popularity of video sensor networks and video celluar phones has generated the need for low-complexity and power-efficient multimedia systems that can handle multiple video input and output streams. While standard video coding techniques fail to satisfy these requirements, distributed source coding is a promising technique for ??uplink?? applications. Wyner-Ziv coding refers to lossy source coding with side information at the decoder. Based on recent theoretical result on successive Wyner-Ziv coding, we propose in this thesis a practical layered Wyner-Ziv video codec using the DCT, nested scalar quantizer, and irregular LDPC code based Slepian-Wolf coding (or lossless source coding with side information) for noiseless channel. The DCT is applied as an approximation to the conditional KLT, which makes the components of the transformed block conditionally independent given the side information. NSQ is a binning scheme that facilitates layered bit-plane coding of the bin indices while reducing the bit rate. LDPC code based Slepian-Wolf coding exploits the correlation between the quantized version of the source and the side information to achieve further compression. Different from previous works, an attractive feature of our proposed system is that video encoding is done only once but decoding allowed at many lower bit rates without quality loss. For Wyner-Ziv coding over discrete noisy channels, we present a Wyner-Ziv video codec using IRA codes for Slepian-Wolf coding based on the idea of two equivalent channels. For video streaming applications where the channel is packet based, we apply unequal error protection scheme to the embedded Wyner-Ziv coded video stream to find the optimal source-channel coding trade-off for a target transmission rate over packet erasure channel.

Thesis (Ph. D.)--University of Washington, 2003. / Vita. Includes bibliographical references (leaves 112-118).

The demand for applications like transmission and sharing of video is ever-increasing. Although network resources (bandwidth in particular) and coverage, networking technologies, compression ratio of state-of-the-art video coders have improved, unreliability of the transmission medium prevents us from gaining the most benefit from these applications. This thesis introduces a video coder that is resilient to network failures for transmission applications by using the framework of multiple description coding (MDC). Unlike traditional video coding which compresses the video into single bitstream, in MDC the video is compressed into more than one bitstream which can be independently decoded. It not only averages out the effect of network errors over the bitstreams but it also makes it possible to utilize the multipath nature of most network topologies. An end-to-end rate-distortion optimization is proposed for the codec to make sure that the codec exhibits improved compression performance and that the descriptions are equally efficient to improve the final video quality. An optimized strategy for packetizing the compressed bitstreams of the descriptions is also proposed which guarantees that each packet is self-contained and efficient. The evaluation of the developed MD codec over simulated unreliable packet networks shows that it is possible to achieve improved resilience with the proposed strategies and the end video quality is significantly improved as a result. This is further verified with subjective evaluation over a range of different types of video test sequences.

Thesis (M.S.)--University of Florida, 2002. / Title from title page of source document. Document formatted into pages; contains vii, 67 p.; also contains graphics. Includes vita. Includes bibliographical references.

Dissertation (Ph.D.)--Worcester Polytechnic Institute. / Keywords: Streaming MPEG, User Study, Video Quality, Forward Error Correction, Temporal Scaling, Quality Scaling. Includes bibliographical references (p.186-198).

Thesis (Ph. D.)--University of Washington, 2005. / Vita. Includes bibliographical references (leaves 93-99).