The Video Detection in the Dynamic Background Based on the Sensitive Areas

Abstract

During the procedure of video analysis with complex dynamic backgrounds, a new method of object detection based on the sensitive areas was presented. First, each frame in a video was converted into an effective information map by using the Harris corner detection method. Second, the sensitive areas in the frame were extracted by using the context information and the effective information maps of the consecutive video frames. The sensitive areas in the video frame were the candidate areas where the target objects would appear at high probabilities. Third, the information entropy features of each sensitive area were extracted to form the feature matrix, based on which, an SVM model was trained for selecting the target areas from the sensitive areas. Finally, the locations of the objects were detected based on the target areas in the video with a complex dynamic background. The experimental results showed that this method could achieve good results against the benchmark of CDnet 2014 on the premise of saving computing resources.