Building a feature-space for visual surveillance

Visual surveillance applications play essential roles as a tool for archiving and to some extend preventing criminal activities. The future of visual surveillance applications relies on the developments in the area of computer vision, human motion analysis and computer algorithms. This paper considers objects low level feature selection and extraction from video frames in order to construct a feature space for visual surveillance applications. Moreover, our paper explains the basic concept of feature extraction and the procedures required to extract these features. Constructing the feature space for surveillance applications is also touched through this paper based on a straightforward scenario of a single human walking. Finally, the paper illustrates the results of the extracted features.