Springer, 2002. — 155 p.In recent years, video has become ubiquitous in daily life. The VCR has become one of the most widely used appliances, typically for recording television programs. Compact video cameras for home use are also very common. Although it has become easier to record video with such technologies, however, editing video can still be difficult or tedious for the average person (despite the development of editing software tools for the personal computer). As a result, VCR and video camera users may amass large quantities of raw, unedited footage that is seldom watched, because segments of interest to the user cannot be easily accessed. In another application, video cameras are also commonly used for surveillance of offices, shops, and homes, but these video streams may require continuous monitoring by security personnel, thus consuming valuable human resources and being prone to lapses of attention by the human observer. These examples suggest that there is a need for automatic analysis of the content of video footage (for example, to facilitate editing, retrieval, or monitoring). Such automation, ideally in real-time, would reduce the burden on the user and broaden the possible applications of video. To pursue such goals, it is useful to take approaches from the field of Computer Vision, one of the most active areas of computer science, that develops algorithms to automatically analyze images acquired by cameras. Recent technical developments have enabled computer vision to deal with video sequences. Such computer vision based video analysis technologies will likely be utilized for a variety of applications, such as telecommunication, video compression, surveillance and security, advanced video games, indexing and retrieval of multimedia database systems, producing digital cinema, and editing video libraries. This book focuses on humans as the subjects of video sequences. This focus is a natural consequence of the immensely important and meaningful role images of people play in daily life. Technically, video sequences of humans are a challenging target for computer vision algorithms, due to the following reasons.Introduction Tracking multiple persons from multiple camera images Posture estimation Recognizing human behavior using Hidden Markov Models Conclusion and Future Work
Чтобы скачать этот файл зарегистрируйтесь и/или войдите на сайт используя форму сверху.