Video analytics for activity recognition in indoor environment using fisheye cameras

Abstract

The analysis of video sequences in order to recognize human activities may assist disabled or elder persons in performing every day activities and detect abnormal situations such as a fall or long periods of inactivity. In this paper we summarize our work concerning the various processing tasks in the case of videos captured by fisheye cameras. The paper includes technical details of the proposed solutions for foreground activity segmentation, human silhouette refinement, camera modeling and activity recognition. The reported results have proven the feasibility and the efficiency of the presented solutions.