Thèses sur le sujet « Video quality prediction »

Transmission of video content over wireless access networks (in particular, Wireless Local Area Networks (WLAN) and Third Generation Universal Mobile Telecommunication System (3G UMTS)) is growing exponentially and gaining popularity, and is predicted to expose new revenue streams for mobile network operators. However, the success of these video applications over wireless access networks very much depend on meeting the user’s Quality of Service (QoS) requirements. Thus, it is highly desirable to be able to predict and, if appropriate, to control video quality to meet user’s QoS requirements. Video quality is affected by distortions caused by the encoder and the wireless access network. The impact of these distortions is content dependent, but this feature has not been widely used in existing video quality prediction models. The main aim of the project is the development of novel and efficient models for video quality prediction in a non-intrusive way for low bitrate and resolution videos and to demonstrate their application in QoS-driven adaptation schemes for mobile video streaming applications. This led to five main contributions of the thesis as follows:(1) A thorough understanding of the relationships between video quality, wireless access network (UMTS and WLAN) parameters (e.g. packet/block loss, mean burst length and link bandwidth), encoder parameters (e.g. sender bitrate, frame rate) and content type is provided. An understanding of the relationships and interactions between them and their impact on video quality is important as it provides a basis for the development of non-intrusive video quality prediction models.(2) A new content classification method was proposed based on statistical tools as content type was found to be the most important parameter. (3) Efficient regression-based and artificial neural network-based learning models were developed for video quality prediction over WLAN and UMTS access networks. The models are light weight (can be implemented in real time monitoring), provide a measure for user perceived quality, without time consuming subjective tests. The models have potential applications in several other areas, including QoS control and optimization in network planning and content provisioning for network/service providers.(4) The applications of the proposed regression-based models were investigated in (i) optimization of content provisioning and network resource utilization and (ii) A new fuzzy sender bitrate adaptation scheme was presented at the sender side over WLAN and UMTS access networks. (5) Finally, Internet-based subjective tests that captured distortions caused by the encoder and the wireless access network for different types of contents were designed. The database of subjective results has been made available to research community as there is a lack of subjective video quality assessment databases.

Context: The newest standard in video coding High Efficiency Video Coding (HEVC) should have an appropriate coder to fully use its potential. There are a lot of video quality assessment methods. These methods are necessary to establish the quality of the video. Objectives: This thesis is a comparison of video quality assessment methods. Objective is to find out which objective method is the most similar to the subjective method. Videos used in tests are encoded in the H.265/HEVC standard. Methods: For testing MSE, PSNR, SSIM methods there is special software created in MATLAB. For VQM method downloaded software was used for testing. Results and conclusions: For videos watched on mobile device: PSNR is the most similar to subjective metric. However for videos watched on television screen: VQM is the most similar to subjective metric. Keywords: Video Quality Assessment, Video Quality Prediction, Video Compression, Video Quality Metrics

The recently released High Efficiency Video Coding (HEVC) standard, which halves the transmission bandwidth requirement of encoded video for almost the same quality when compared to H.264/AVC, and the availability of increased network bandwidth (e.g. from 2 Mbps for 3G networks to almost 100 Mbps for 4G/LTE) have led to the proliferation of video streaming services. Based on these major innovations, the prevalence and diversity of video application are set to increase over the coming years. However, the popularity and success of current and future video applications will depend on the perceived quality of experience (QoE) of end users. How to measure or predict the QoE of delivered services becomes an important and inevitable task for both service and network providers. Video quality can be measured either subjectively or objectively. Subjective quality measurement is the most reliable method of determining the quality of multimedia applications because of its direct link to users’ experience. However, this approach is time consuming and expensive and hence the need for an objective method that can produce results that are comparable with those of subjective testing. In general, video quality is impacted by impairments caused by the encoder and the transmission network. However, videos encoded and transmitted over an error-prone network have different quality measurements even under the same encoder setting and network quality of service (NQoS). This indicates that, in addition to encoder settings and network impairment, there may be other key parameters that impact video quality. In this project, it is hypothesised that video content type is one of the key parameters that may impact the quality of streamed videos. Based on this assertion, parameters related to video content type are extracted and used to develop a single metric that quantifies the content type of different video sequences. The proposed content type metric is then used together with encoding parameter settings and NQoS to develop content-based video quality models that estimate the quality of different video sequences delivered over IP-based network. This project led to the following main contributions: (1) A new metric for quantifying video content type based on the spatiotemporal features extracted from the encoded bitstream. (2) The development of novel subjective test approach for video streaming services. (3) New content-based video quality prediction models for predicting the QoE of video sequences delivered over IP-based networks. The models have been evaluated using subjective and objective methods.

Along with the rapid growth in consumer adoption of modern portable devices, video streaming is expected to dominate a large share of the global Internet traffic in the near future. Today user experience is becoming a reliable indicator for video service providers and telecommunication operators to convey overall end-to-end system functioning. Towards this, there is a profound need for an efficient Quality of Experience (QoE) monitoring and prediction. QoE is a subjective metric, which deals with user perception and can vary due to the user expectation and context. However, available QoE measurement techniques that adopt a full reference method are impractical in real-time transmission since they require the original video sequence to be available at the receiver’s end. QoE prediction, however, requires a firm understanding of those Quality of Service (QoS) factors that are the most influential on QoE. The main aim of this thesis work is the development of novel and efficient models for video quality prediction in a non-intrusive way and to demonstrate their application in QoE-enabled optimisation schemes for video delivery. In this thesis, the correlation between QoS and QoE is utilized to objectively estimate the QoE. For this, both objective and subjective methods were used to create datasets that represent the correlation between QoS parameters and measured QoE. Firstly, the impact of selected QoS parameters from both encoding and network levels on video QoE is investigated. The obtained QoS/QoE correlation is backed by thorough statistical analysis. Secondly, the development of two novel hybrid non-reference models for predicting video quality using fuzzy logic inference systems (FIS) as a learning-based technique. Finally, attention was move onto demonstrating two applications of the developed FIS prediction model to show how QoE is used to optimise video delivery.

Technological advances in telecommunications and computing power have driven a massive increase in the availability of digital video, both in the traditional broadcast environment and in on-demand scenarios such as the internet and IPTV. In order to support this increase video compression technologies have developed both to reduce bitrates but also to utilise the additional computing power available. The next step for video compression is commonly thought to be in exploitation of the Human Visual System (HVS) in perceptual coding, although this continues to suffer from the difficulty of evaluating the quality of compressed video. This thesis proposes a perceptual video compression framework - from quality estimation, through gaze estimation to variable quality coding.

When a demanding video decoding task requires more CPU resources then available, playback degrades ungracefully today: The decoder skips frames selected arbitrarily or by simple heuristics, which is noticed by the viewer as jerky motion in the good case or as images completely breaking up in the bad case. The latter can happen due to missing reference frames. This thesis provides a way to schedule individual decoding tasks based on a cost for performance trade. Therefore, I will present a way to preprocess a video, generating estimates for the cost in terms of execution time and the performance in terms of perceived visual quality. The granularity of the scheduling decision is a single slice, which leads to a much more fine-grained approach than dealing with entire frames. Together with an actual scheduler implementation that uses the generated estimates, this work allows for higher perceived quality video playback in case of CPU overload
Wenn eine anspruchsvolle Video-Dekodierung mehr Prozessor-Ressourcen benötigt, als verfügbar sind, dann verschlechtert sich die Abspielqualität mit aktuellen Methoden drastisch: Willkürlich oder mit einfachen Heuristiken ausgewählten Bilder werden nicht dekodiert. Diese Auslassung nimmt der Betrachter im günstigsten Fall nur als ruckelnde Bewegung wahr, im ungünstigen Fall jedoch als komplettes Zusammenbrechen nachfolgender Bilder durch Folgefehler im Dekodierprozess. Meine Arbeit ermöglicht es, einzelne Teilaufgaben des Dekodierprozesses anhand einer Kosten-Nutzen-Analyse einzuplanen. Dafür ermittle ich die Kosten im Sinne von Rechenzeitbedarf und den Nutzen im Sinne von visueller Qualität für einzelne Slices eines H.264 Videos. Zusammen mit einer Implementierung eines Schedulers, der diese Werte nutzt, erlaubt meine Arbeit höhere vom Betrachter wahrgenommene Videoqualität bei knapper Prozessorzeit

Video streaming applications over best-effort networks, such as the Internet, have become very popular among Internet users. Watching live sports and news, renting movies, watching clips online, making video calls, and participating in videoconferences are typical video applications that millions of people use daily. One of the most challenging aspects of video communication is the proper transmission of video in various network bandwidth conditions. Currently, various devices with different processing powers and various connection speeds (2G, 3G, Wi-Fi, and LTE) are used to access video over the Internet, which offers best-effort services only. Skype, ooVoo, Yahoo Messenger, and Zoom are some well-known applications employed on a daily basis by people throughout the world; however, best-effort networks are characterized by dynamic and unpredictable changes in the available bandwidth, which adversely affect the quality of the video. For the average consumer, there is no guarantee of receiving an exact amount of bandwidth for sending or receiving video data. Therefore, the video delivery system must use a bandwidth adaptation mechanism to deliver video content properly. Otherwise, bandwidth variations will lead to degradation in video quality or, in the worst case, disrupt the entire service. This is especially problematic for videoconferencing (VC) because of the bulkiness of the video, the stringent bandwidth demands, and the delay constraints. Furthermore, for business grade VC, which uses high definition videoconferencing (HDVC), user expectations regarding video quality are much higher than they are for ordinary VC. To manage network fluctuations and handle the video traffic, two major components in the system should be improved: the video encoder and the congestion control. The video encoder is responsible for compressing raw video captured by a camera and generating a bitstream. In addition to the efficiency of the encoder and compression speed, its output flow is also important. Though the nature of video content may make it impossible to generate a constant bitstream for a long period of time, the encoder must generate a flow around the given bitrate. While the encoder generates the video traffic around the given bitrate, congestion management plays a key role in determining the current available bandwidth. This can be done by analyzing the statistics of the sent/received packets, applying mathematical models, updating parameters, and informing the encoder. The performance of the whole system is related to the in-line collaboration of the encoder and the congestion management, in which the congestion control system detects and calculates the available bandwidth for a specific period of time, preferably per incoming packet, and informs rate control (RC) to adapt its bitrate in a reasonable time frame, so that the network oscillations do not affect the perceived quality on the decoder side and do not impose adverse effects on the video session. To address these problems, this thesis proposes a collaborative management architecture that monitors the network situation and manages the encoded video rate. The goal of this architecture is twofold: First, it aims to monitor the available network bandwidth, to predict network behavior and to pass that information to the encoder. So encoder can encode a suitable video bitrate. Second, by using a smart rate controller, it aims for an optimal adaptation of the encoder output bitrate to the bitrate determined by congestion control. Merging RC operations and network congestion management, to provide a reliable infrastructure for HDVC over the Internet, represents a unique approach. The primary motivation behind this project is that by applying videoconference features, which are explained in the rate controller and congestion management chapter, the HDVC application becomes feasible and reliable for the business grade application even in the best-effort networks such as the Internet.

Video Image Analysis (VIA) is a digital camera based technology that extracts relevant information from images using purpose tailored image processing software. In the present work, the VSS2000 image analysis system from E+V Technology GmbH has been used in a large lamb abattoir to determine the value of carcasses in an objective, consistent and automated way. In this thesis results are reported of several experiments conducted within the framework of two UK-funded projects. The aims of the research were (i) the calibration and validation of the VIAtechnique for the evaluation of lamb carcasses under UK abattoir conditions, with the view to scientifically examine the accuracy and precision of information from the VIA systems as the basis for a value-based marketing system, (ii) to investigate the use of VIA measurements (weights of primal meat yields and carcass dimensional measurements) in sheep breeding programmes to improve carcass and meat quality and (iii) to evaluate the potential of this technology to reward increased carcass quality associated with the use of breeding strategies based on the inclusion of a quantitative trait locus (QTL) for improved muscularity. Accuracy, precision and consistency of The Meat and Livestock Commission (MLC) carcass classification scheme, currently used in UK abattoirs to evaluate carcass quality, was compared against the VIA system in the prediction of various primal joint weights. The results highlighted the advantage of the VIA system being on average 2% more accurate (measured as coefficient of determination: R2) and 12% more precise (measured as root meat squared error: RMSE) in predicting weight of primal meat yields (leg, chump, loin, breast and shoulder) of the lamb carcasses than the MLC carcass classification scheme. The genetic analysis of VIA-based predicted primal joint weights showed substantial additive genetic variance, suggesting that their use in sheep breeding programmes could improve carcass quality either by an improvement of conformation or by an increased weight of the most valuable primal cuts, without an increase in fatness. Favourable associations between VIA primal weights and performance traits indicate that selection based on VIA traits is possible without a negative effect on average daily gain, live weight and cold carcass weight. Although computer tomography (CT) and dissection found in related studies significant effects of a Texel muscling-QTL (TM-QTL) for increased muscularity in the loin region, in the present study they could not be identified by both, the current industry carcass evaluation system for conformation and fatness and the VIA system. A calibration of the VIA system against CT measurements resulted in improved VIA prediction equations for primal meat yields and also showed a moderate potential to estimate loin muscle traits measured by CT and to detect partially the effect of the TM-QTL on these traits. The results of the research demonstrated that VIA is a consistent method to measure carcass composition and that it improved the prediction (accuracy and precision) of primal meat yields compared to the present MLC scoring system. The estimated genetic parameters for VIA primal meat yields suggested that selection for increased lean meat yield from lamb carcass measured using VIA can contribute to genetic improvement of carcass quality without increasing carcass fatness. The results suggest that VIA technology installed in abattoirs could provide the means for the development of a value-based marketing system by paying for weights of the most valuable primal cuts measured using VIA.

Quality of Experience (QoE) is the degree of delight or annoyance of the user of an application or service [1]. To ensure a proper level of QoE for end users, networks and service providers have to continuously monitor their systems in terms of technical parameters, which can then be used to estimate QoE. Especially for video streaming services, which consume a large amount of traffic, network problems such as bandwidth fluctuations quickly develop into annoying artefacts visible to the users, which may lead to abandonment of services. Internet Service Providers (ISPs) are therefore continuously monitoring video network streams in order to provide the better QoE. In this regard to conduct the user behavioral studies, the ISPs spend a large amount of money and energy every time. To avoid this, we are using existing user behavioral studies and simulating the user behavior in an automated set-up and try to measure the impact of network conditions. In our current studies based on the user-behavioral model used [5], we can conclude that low upload speeds don’t affect on simulated user behavior unless they are in high download speed networks. Simulated users with the mid-range download and upload bandwidth tend to face more stalling and quality switches compared to both low and high-bandwidth users. Key quality indicators(KQIs) of video QoE also depends on the number of videos we measure in a single session. Reloading of player helps to reduce stalling for mid and high bandwidths. Reloading worsens the situation in low bandwidth scenarios.
Kvalitet av erfarenhet (QoE) definieras som: "Graden av fröjd eller förargelse av användaren av en applikation eller en service. Den resulterar från uppfyllelsen av hans eller hennes förväntningar med hänsyn till hjälpmedlet, och/eller njutning av applikationen eller servicen i ljuset av användarens personlighet och aktuella tillstånd" [1]. Att att se till en riktig nivå av QoE för slut användare, nätverk och tjänstefamiljeförsörjare måste övervakar fortlöpande deras system när det gäller tekniska parametrar, som kan därefter vara den van vid bedömningen QoE. Speciellt för videoen som strömmar service, som konsumerar ett stort belopp av trafik, framkallar nätverksproblem liksom bandbreddväxlingar snabbt in i förargliga artefacts som är synliga till användarena, som kan leda till övergivande av service. Internetleverantörer (ISPs) är därför fortlöpande videopp nätverksströmmar för övervakning för att att ge den bättre QoEen. I detta avseende att föra de beteendestudierna för användaren, spenderar ISPsna en stor mängd pengar och energi varje gång. Att undvika denna, använder simulerar vi beteendestudier för existerande användare och användareuppförandet i en automatiserat aktivering och försök för att mäta inverkan av nätverksvillkor. I våra aktuella studier som baseras på den använda användare-beteendemodellen [5], oss kan avsluta som laddar upp lågt hastigheter inte påverkar på simulerat användareuppförande, om inte de är i höga nedladdninghastighetsnätverk. Simulerade användare med mitt–området nedladdar och laddar upp bandbredd ansar för att vända mot mer avbrottsoch kvalitetsströmbrytare som jämförs till både låga och höga bandbreddanvändare. Nyckelkvalitetsindikatorer (KQIs) av video QoE beror också på numret av video som vi mäter i en enkel period. Tillbakaläggande av spelaren hjälper att förminska avbrott för mittoch höga bandbredder. Tillbakaläggande försämrar läget i scenarion för låg bandbredd.

Objective Video Quality Assessment plays a vital role in visual processing systems and especially in the mobile communication field, some of the video applications boosted the interest of robust methods for quality assessment. Out of all existing methods for Video Quality Analysis, No-Reference (NR) Video Quality Assessment is the one which is most needed in situations where the handiness of reference video is not available. Our challenge lies in formulating and melding effective features into one model based on human visualizing characteristics. Our research work explores the tradeoffs between quality prediction and complexity of a system. Therefore, we implemented support vector regression algorithm as NR-based Video Quality Metric(VQM) for quality estimation with simplified input features. The features are obtained from extraction of H.264 bitstream data at the decoder side of the network. Our metric predicted with Pearson correlation coefficient of 0.99 for SSIM, 0.98 for PEVQ, 0.96 for subjective score and 0.94 for PSNR metric. Therefore in terms of prediction accuracy, the proposed model has good correlation with all deployed metrics and the obtained results demonstrates the robustness of our approach. In our research work, the proposed metric has a good correlation with subjective scores which concludes that proposed metric can be employed for real time use, since subjective scores are considered as true or standard values of video quality.

Diploma thesis deals with the analysis of quality in digital television systems and contains theoretical description of subjective and objective assessment of quality picture methods. The thesis contains short-term and long-term analysis of quality picture of terrestrial television DVB-T. Measurements and experimentations were carried out with the help of Rohde&Schwarz DVQ analyzer of picture quality and software MPEG-2 Quality Monitor and MPEG-2 Elementary stream analyzer.

Les progrès réalisés en systèmes micro-électro-mécaniques couplés avec leur convergence vers les systèmes de communication sans fil, ont permis l'émergence des réseaux de capteurs sans fil (RCSF). Les contraintes de ces réseaux font que tous les efforts soient fournis pour proposer des solutions économes en énergie. Avec les récents développements des technologies CMOS, des capteurs d'images à faible coût ont été développés. En conséquence, un nouveau dérivé des RCSF, qui sont les Réseaux de Capteurs Vidéo Sans Fil (RCVSF), a été proposé. La particularité des données vidéo ainsi que les contraintes inhérentes aux nœuds ont introduit de nouveaux défis. Dans cette thèse, nous proposons deux solutions basées sur l'approche inter-couches pour la livraison de la vidéo sur les RCVSF. La première solution propose un nouveau schéma de compression vidéo adaptatif, efficace en énergie et basé sur la norme de compression vidéo H.264/AVC. Le flux vidéo est ensuite géré par une version améliorée du protocole MMSPEED que nous proposons et notons EQBSA-MMSPEED. Les résultats des simulations montrent que la durée de vie du réseau est étendue de 33%, tout en améliorant la qualité du flux vidéo reçu de 12%. Dans la deuxième solution, nous enrichissons le schéma de compression de modèles mathématiques pour prévoir la consommation d'énergie et la distorsion de l'image lors des phases d'encodage et de transmission. Le flux vidéo est géré par un nouveau protocole de routage efficace en énergie et à fiabilité améliorée noté ERMM. Comparée à une approche basique, cette solution réalise une extension de la durée de vie du réseau de 15%, tout en améliorant la qualité du flux vidéo reçu de 35%
Thanks 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%

The use of audio and video communication has increased exponentially over the last decade and has gone from speech over GSM to HD resolution video conference between continents on mobile devices. As the use becomes more widespread the interest in delivering high quality media increases even on devices with limited resources. This includes both development and enhancement of the communication chain but also the topic of objective measurements of the perceived quality. The focus of this thesis work has been to perform enhancement within speech encoding and video decoding, to measure influence factors of audio and video performance, and to build methods to predict the perceived video quality. The audio enhancement part of this thesis addresses the well known problem in the GSM system with an interfering signal generated by the switching nature of TDMA cellular telephony. Two different solutions are given to suppress such interference internally in the mobile handset. The first method involves the use of subtractive noise cancellation employing correlators, the second uses a structure of IIR notch filters. Both solutions use control algorithms based on the state of the communication between the mobile handset and the base station. The video enhancement part presents two post-filters. These two filters are designed to improve visual quality of highly compressed video streams from standard, block-based video codecs by combating both blocking and ringing artifacts. The second post-filter also performs sharpening. The third part addresses the problem of measuring audio and video delay as well as skewness between these, also known as synchronization. This method is a black box technique which enables it to be applied on any audiovisual application, proprietary as well as open standards, and can be run on any platform and over any network connectivity. The last part addresses no-reference (NR) bitstream video quality prediction using features extracted from the coded video stream. Several methods have been used and evaluated: Multiple Linear Regression (MLR), Artificial Neural Network (ANN), and Least Square Support Vector Machines (LS-SVM), showing high correlation with both MOS and objective video assessment methods as PSNR and PEVQ. The impact from temporal, spatial and quantization variations on perceptual video quality has also been addressed, together with the trade off between these, and for this purpose a set of locally conducted subjective experiments were performed.

This dissertation considers packet-switched wireless networks for transmission of variable-rate layered hybrid video streams. Target applications are video streaming and broadcasting services. The work can be divided into two main parts.

In the first part, a novel quality-scalable scheme based on coefficient refinement and encoder quality constraints is developed as a possible extension to the video coding standard H.264. After a technical introduction to the coding tools of H.264 with the main focus on error resilience features, various quality scalability schemes in previous research are reviewed. Based on this discussion, an encoder decoder framework is designed for an arbitrary number of quality layers, hereby also enabling region-of-interest coding. After that, the performance of the new system is exhaustively tested, showing that the bit rate increase typically encountered with scalable hybrid coding schemes is, for certain coding parameters, only small to moderate. The double- and triple-layer constellations of the framework are shown to perform superior to other systems.

The second part considers layered code streams as generated by the scheme of the first part. Various error propagation issues in hybrid streams are discussed, which leads to the definition of a decoder quality constraint and a segmentation of the code stream to transmit. A packetization scheme based on successive source rate consumption is drafted, followed by the formulation of the channel code rate optimization problem for an optimum assignment of available codes to the channel packets. Proper MSE-based error metrics are derived, incorporating the properties of the source signal, a terminate-on-error decoding strategy, error concealment, inter-packet dependencies, and the channel conditions. The Viterbi algorithm is presented as a low-complexity solution to the optimization problem, showing a great adaptivity of the joint source channel coding scheme to the channel conditions. An almost constant image qualiity is achieved, also in mismatch situations, while the overall channel code rate decreases only as little as necessary as the channel quality deteriorates. It is further shown that the variance of code distributions is only small, and that the codes are assigned irregularly to all channel packets.

A double-layer constellation of the framework clearly outperforms other schemes with a substantial margin.

Keywords — Digital lossy video compression, visual communication, variable bit rate (VBR), SNR scalability, layered image processing, quality layer, hybrid code stream, predictive coding, progressive bit stream, joint source channel coding, fidelity constraint, channel error robustness, resilience, concealment, packet-switched, mobile and wireless ATM, noisy transmission, packet loss, binary symmetric channel, streaming, broadcasting, satellite and radio links, H.264, MPEG-4 AVC, Viterbi, trellis, unequal error protection

Dissertação de Mestrado em Engenharia Informática apresentada à Faculdade de Ciências e Tecnologia da Universidade de Coimbra
Remote video viewing is largely growing, especially in wireless devices. Therefore, video transmission mechanisms such as retransmission and Forward Error Correction (FEC) are very important to ensure that the video arrives at its destination while retaining its quality. While some of these mechanisms can obtain good performance, this does not mean they provide an optimal situation. There are several factors that negatively impact Quality of Experience (QoE) in video transmission, such as network congestion, loss and type of video being transmitted. Adaptive Forward Error Correction schemes aim at providing protection against loss errors by analysing several relevant characteristics of the video and/or network, and using them to apply unequal error protection (UEP) to di erent parts of the video. Taking this into account, it is clear that an adaptive FEC mechanism is a viable option to achieve optimal error protection, thus increasing the video transmission's resilience against errors and losses, and therefore improving Quality of Experience for the end user. This work presents and examines the state-of-the-art mechanisms for retransmission, FEC, QoE, video classi cation and algorithm optimization. The assessment of the mechanisms presents an in-depth study of each concept and individually explores their advantages and disadvantages. Three mechanisms which provide adaptive FEC are developed in and evaluated in a simulation environment, with motion classi cation, error correction, adaptive redundancy allocation and loss feedback components. The capabilities of the adaptive FEC mechanisms are evaluated through a wireless network environment in order to assess their performance against other methods. The results obtained through the assessed mechanisms showed incremental improvements through each stage of development. They also showed that it is possible to accurately characterize the intensity of motion in video sequences according to their characteristics. Furthermore, the ndings highlight the importance of shielding sequences with higher amounts of motion intensity with greater quantities of redundancy due to david their higher degradation when subjected to error. The mechanisms achieved a considerable reduction in the amount of redundancy used to shield the data being transmitted, particularly the loss prediction mechanism. At the same time, the mechanisms maintained video quality which combined with the overhead reduction results in an overall improvement of QoE ultimately enhancing video transmission in wireless networks.

When a demanding video decoding task requires more CPU resources then available, playback degrades ungracefully today: The decoder skips frames selected arbitrarily or by simple heuristics, which is noticed by the viewer as jerky motion in the good case or as images completely breaking up in the bad case. The latter can happen due to missing reference frames. This thesis provides a way to schedule individual decoding tasks based on a cost for performance trade. Therefore, I will present a way to preprocess a video, generating estimates for the cost in terms of execution time and the performance in terms of perceived visual quality. The granularity of the scheduling decision is a single slice, which leads to a much more fine-grained approach than dealing with entire frames. Together with an actual scheduler implementation that uses the generated estimates, this work allows for higher perceived quality video playback in case of CPU overload.
Wenn eine anspruchsvolle Video-Dekodierung mehr Prozessor-Ressourcen benötigt, als verfügbar sind, dann verschlechtert sich die Abspielqualität mit aktuellen Methoden drastisch: Willkürlich oder mit einfachen Heuristiken ausgewählten Bilder werden nicht dekodiert. Diese Auslassung nimmt der Betrachter im günstigsten Fall nur als ruckelnde Bewegung wahr, im ungünstigen Fall jedoch als komplettes Zusammenbrechen nachfolgender Bilder durch Folgefehler im Dekodierprozess. Meine Arbeit ermöglicht es, einzelne Teilaufgaben des Dekodierprozesses anhand einer Kosten-Nutzen-Analyse einzuplanen. Dafür ermittle ich die Kosten im Sinne von Rechenzeitbedarf und den Nutzen im Sinne von visueller Qualität für einzelne Slices eines H.264 Videos. Zusammen mit einer Implementierung eines Schedulers, der diese Werte nutzt, erlaubt meine Arbeit höhere vom Betrachter wahrgenommene Videoqualität bei knapper Prozessorzeit.