Human activity can be described as a sequence of 3D body postures. The traditional approach to recognition and 3D reconstruction of human activity has been to track motion in 3D, m...
The context-centered approach to object detection and recognition is based on the intuition that the contextual information of real-world scenes provides relevant information for ...
Pattern recognition methods for complex structured objects such as handwritten characters often have to deal with vast search spaces. Developed techniques, despite significant adv...
Many computer vision systems try to infer semantic information about a video scene content by looking at the time series of the silhouettes of the moving objects. This paper propo...
This paper proposes a method of evolutionary robot vision based on a steady-state genetic algorithm and fuzzy evaluation. In order to improve the communication capability of human-...