Multiple-human tracking using multiple cameras

摘要: We propose a human motion detection method using multiple-viewpoint images. In vision-based tracking, self-occlusions and human-human occlusions are part of the more significant problems. Employing multiple viewpoints viewpoint selection mechanism, however, can reduce these The vision system in this case should select best for extracting information; "best" selections be changed among different types target information. address problem tracking bodies. divide task into three primitive sub-tasks (position detection, rotation angle body side detection). Each sub-task has criterion selecting an estimation result one help another sub-task. describe criteria accomplishing individual relationships between sub-tasks. have built experimental based on small number reliable image features performed fundamental examinations approach.