Uma Arquitetura de Hardware para estimação de movimento aplicada à compressão de vídeo digital

A tarefa de estimação de movimento, utilizada na compressão de vídeo digital, é normalmente realizada em hardware por processador dedicado, uma vez que demanda expressiva capacidade computacional. Este trabalho propõe e desenvolve uma arquitetura de hardware para realizar o cálculo dos vetores de movimento no contexto de compressão de vídeo digital. Essa arquitetura para estimação de movimento é composta pelos blocos: interface de entrada e saída (E/S), matriz de processamento com 64 elementos de processamento, unidade de comparação e unidade de controle. A arquitetura foi descrita em linguagem VHDL de maneira que o número de bits utilizados para representação da luminância dos pontos é configurável. A partir desta descrição, foi gerado um protótipo para dados representados em 4 bits utilizando um kit de desenvolvimento baseado no dispositivo FPGA XC2S150 da Xilinx. Para validação do algoritmo e da arquitetura implementada, além da simulação, foi desenvolvido um software para plataforma PC capaz de exercitar as funcionalidades do protótipo. O PC é utilizado como dispositivo controlador de E/S para esta validação, na qual uma implementação do algoritmo em software e outra em linguagem de descrição de hardware são comparadas. A máxima freqüência de trabalho do protótipo, estimada por simulação da arquitetura mapeada no FPGA XC2S150, é de 33 MHz. A esta freqüência o núcleo da arquitetura paralela de 64 elementos de processamento realiza cerca de 2,1 GOps (bilhões de operações inteiras por segundo). Esta arquitetura de hardware calcula os vetores de movimento para vídeo no formato 640x480 pontos à taxa de 107,32 quadros por segundo, ou um quadro a cada 9,3 ms. A arquitetura implementada para luminânica em 4 bits ocupa 16 pinos de E/S, 71,1% dos blocos lógicos do FPGA e 83,3% dos blocos de memória disponíveis no dispositivo XC2S150.

MoCHA : arquitetura dedicada para a compensação de movimento em decodificadores de vídeo de alta definição, seguindo o padrão H.264

O padrão H.264 foi desenvolvido pelo JVT, que foi formado a partir de uma união entre os especialistas do VCEG da ITU-T e do MPEG da ISO/IEC. O padrão H.264 atingiu seu objetivo de alcançar as mais elevadas taxas de processamento dentre todos os padrões existentes, mas à custa de um grande aumento na complexidade computacional. Este aumento de complexidade impede, pelo menos na tecnologia atual, a utilização de codecs H.264 implementados em software, quando se deseja a decodi cação de vídeos de alta de nição em tempo real. Essa dissertação propõe uma solução arquitetural de hardware, denominada MoCHA, para compensação de movimento do decodi cador de vídeo de alta de nição, segundo o padrão H.264/AVC. A MoCHA está dividida em três blocos principais, a predição dos vetores de movimento, o acesso à memória e o processamento de amostras. A utilização de uma cache para explorar a redundância dos dados nos acessos à mem ória, em conjunto com melhorias propostas, alcançou economia de acessos à memória superior a 60%, para os casos testados. Quando uma penalidade de um ciclo por troca de linha de memória é imposta, a economia de ciclos de acesso supera os 75%. No processamento de amostras, a arquitetura realiza o processamento dos dois blocos, que dão origem ao bloco bi-preditivo, de forma serial. Dessa forma, são economizados recursos de hardware, uma vez que a duplicação da estrutura de processamento não é requerida. A arquitetura foi validada a partir de simulações, utilizando entradas extraídas de seqüências codi cadas. Os dados extraídos, salvos em arquivos, serviam de entrada para a simulação. Os resultados da simulação foram salvos em arquivos e comparados com os resultados extraídos. O processador de amostras do compensador de movimento foi prototipado na placa XUP Virtex-II Pro. A placa possui um FPGA VP30 da família Virtex-II PRO da Xilinx. O processador PowerPC 405, presente no dispositivo, foi usado para implementar um test bench para validar a operação do processador de amostras mapeado para o FPGA. O compensador de movimento para o decodi cador de vídeo H.264 foi descrito em VHDL, num total de 30 arquivos e cerca de 13.500 linhas de código. A descrição foi sintetizada pelo sintetizador Syplify Pro da Symplicity para o dispositivo XC2VP30-7 da Xilinx, consumindo 8.465 slices, 5.671 registradores, 10.835 LUTs, 21 blocos de memó- ria interna e 12 multiplicadores. A latência mínima para processar um macrobloco é de 233 ciclos, enquanto a máxima é de 590, sem considerar misses na cache. A freqüência máxima de operação foi de 100,5 MHz. A arquitetura projetada é capaz de processar, no pior caso, 36,7 quadros HDTV de 1080 por 1920, inteiramente bi-preditivos, por segundo. Para quadros do tipo P, que não utilizam a bi-predição, a capacidade de processamento sobe para 64,3 quadros por segundo. A arquitetura apresentada para o processamento de quadros bi-preditivos e a hierarquia de memória são, até o momento, inéditas na literatura. Os trabalhos relativos a decodi cadores completos não apresentam a solução para esse processamento. Os resultados apresentados tornam a MoCHA uma solução arquitetural capaz de fazer parte de um decodi cador para vídeos de alta definição.

Talking about watching : using the video card game and wiki-web technology to engage IT students in developing observational skills.

Designers need to develop good observational skills in order to conduct user studies that reveal the subtleties of human interactions and adequately inform design activity. In this paper we describe a game format that we have used in concert with wiki-web technology, to engage our IT and Information Environments students in developing much sharper observational skills. The Video Card Game is a method of video analysis that is suited to design practitioners as well as to researchers. It uses the familiar format of a card game similar to "Happy Families,, to help students develop themes of interactions from watching video clips. Students then post their interaction themes on wiki-web pages, which allows the teaching team and other students to edit and comment on them. We found that the tangible (cards), game, role playing and sharing aspects of this method led to a much larger amount of interaction and discussion between student groups and between students and the teaching team, than we have achieved using our traditional teaching methods, while taking no more time on the part of the teaching staff. The quality of the resulting interaction themes indicates that this method fosters development of observational skills.In the paper we describe the motivations, method and results in full. We also describe the research context in which we collected the videotape data, and how this method relates to state of the art research methods in interaction design for ubiquitous computing technology.

Lawful withdrawal: The video

When should a person who has a heart attack not be resuscitated? When should a patient no longer be kept alive on a ventilator, or be provided with food and water by a tube? When should a person not be given a blood transfusion they need to stay alive? The answers to these questions depend on a number of factors including the mental or physical condition of the patient and any wishes they have expressed prior to losing the ability to make this decision, as well as the requirements of good medical practice. This video is a record of a public lecture held on 7 July 2004 by the Faculty of Law at the Queensland University of Technology, in association with the Faculty of Health, the Centre for Palliative Care Research and Education, and Palliative Care Queensland.

Brand placements in computer and video games : an overview and research questions

Despite the size and growth of the computer and video gaming industry – as well as the increasing use of the medium for the placement of advertising and product placement – researchers have neglected this area. By drawing on existing literature and research in similar and related areas of film product placement, sponsorship and interactivity, the authors present a conceptual overview and identify areas for research.

A review of the video games industry in the United States

With increasing revenues for video game manufacturers, higher software sales and a more diverse audience, the video games industry has been experiencing strong and rapid growth in recent times, rivalling other forms of entertainment. As a result, games have begun to attract the attention of marketing practitioners who are finding it increasingly difficult to attract consumer attention, and are seeking alternative media for marketing communications. This paper provides a review of the video games industry in the United States and raises the question as to whether games are a viable new medium for marketing messages. Areas for research are identified.

Exploring product placement in video games : an investigation of recall effects

Despite its growth and prominence, product placement is generally under-researched and this is even more apparent in the area of placement in video gaming. This paper presents exploratory focus group research into this practice. Findings indicate that the introductory footage to a game provides placement opportunities with the highest level of recall, while peripheral non-action is the worst. Interestingly, recall also appears to be higher for individual brands as opposed to manufacturer brands.

Digital media literacy : connecting young people’s identities, creative production and learning about video games

This chapter outlines examples of classroom activities that aim to make connections between young people’s everyday experiences with video games and the formal high school curriculum. These classroom activities were developed within the emerging field of digital media literacy. Digital media literacy combines elements of ‘traditional’ approaches to media education with elements of technology and information education (Buckingham, 2007; Warschauer, 2006). It is an educational field that has gained significant attention in recent years. For example, digital media literacy has become a significant objective for media policy makers in response to the increased social and cultural roles of new media technologies and controversies associated with young people’s largely unregulated online participation. Media regulators, educational institutions and independent organizations1 in the United States, Canada, the United Kingdom and Australia have developed digital media literacy initiatives that aim to provide advice to parents, teachers and young people.

Improved detection and tracking of objects in surveillance video

Surveillance networks are typically monitored by a few people, viewing several monitors displaying the camera feeds. It is then very difficult for a human operator to effectively detect events as they happen. Recently, computer vision research has begun to address ways to automatically process some of this data, to assist human operators. Object tracking, event recognition, crowd analysis and human identification at a distance are being pursued as a means to aid human operators and improve the security of areas such as transport hubs. The task of object tracking is key to the effective use of more advanced technologies. To recognize an event people and objects must be tracked. Tracking also enhances the performance of tasks such as crowd analysis or human identification. Before an object can be tracked, it must be detected. Motion segmentation techniques, widely employed in tracking systems, produce a binary image in which objects can be located. However, these techniques are prone to errors caused by shadows and lighting changes. Detection routines often fail, either due to erroneous motion caused by noise and lighting effects, or due to the detection routines being unable to split occluded regions into their component objects. Particle filters can be used as a self contained tracking system, and make it unnecessary for the task of detection to be carried out separately except for an initial (often manual) detection to initialise the filter. Particle filters use one or more extracted features to evaluate the likelihood of an object existing at a given point each frame. Such systems however do not easily allow for multiple objects to be tracked robustly, and do not explicitly maintain the identity of tracked objects. This dissertation investigates improvements to the performance of object tracking algorithms through improved motion segmentation and the use of a particle filter. A novel hybrid motion segmentation / optical flow algorithm, capable of simultaneously extracting multiple layers of foreground and optical flow in surveillance video frames is proposed. The algorithm is shown to perform well in the presence of adverse lighting conditions, and the optical flow is capable of extracting a moving object. The proposed algorithm is integrated within a tracking system and evaluated using the ETISEO (Evaluation du Traitement et de lInterpretation de Sequences vidEO - Evaluation for video understanding) database, and significant improvement in detection and tracking performance is demonstrated when compared to a baseline system. A Scalable Condensation Filter (SCF), a particle filter designed to work within an existing tracking system, is also developed. The creation and deletion of modes and maintenance of identity is handled by the underlying tracking system; and the tracking system is able to benefit from the improved performance in uncertain conditions arising from occlusion and noise provided by a particle filter. The system is evaluated using the ETISEO database. The dissertation then investigates fusion schemes for multi-spectral tracking systems. Four fusion schemes for combining a thermal and visual colour modality are evaluated using the OTCBVS (Object Tracking and Classification in and Beyond the Visible Spectrum) database. It is shown that a middle fusion scheme yields the best results and demonstrates a significant improvement in performance when compared to a system using either mode individually. Findings from the thesis contribute to improve the performance of semi-automated video processing and therefore improve security in areas under surveillance.

Robust video stabilisation algorithm using feature point selection and delta optical flow

In this study, the authors propose a novel video stabilisation algorithm for mobile platforms with moving objects in the scene. The quality of videos obtained from mobile platforms, such as unmanned airborne vehicles, suffers from jitter caused by several factors. In order to remove this undesired jitter, the accurate estimation of global motion is essential. However it is difficult to estimate global motions accurately from mobile platforms due to increased estimation errors and noises. Additionally, large moving objects in the video scenes contribute to the estimation errors. Currently, only very few motion estimation algorithms have been developed for video scenes collected from mobile platforms, and this paper shows that these algorithms fail when there are large moving objects in the scene. In this study, a theoretical proof is provided which demonstrates that the use of delta optical flow can improve the robustness of video stabilisation in the presence of large moving objects in the scene. The authors also propose to use sorted arrays of local motions and the selection of feature points to separate outliers from inliers. The proposed algorithm is tested over six video sequences, collected from one fixed platform, four mobile platforms and one synthetic video, of which three contain large moving objects. Experiments show our proposed algorithm performs well to all these video sequences.