Constant execution time multiple human detectors regardless of target number increase based on CHLAC

This paper proposes the constant execution time multiple human detection system. The human detection from video sequence is important in a monitoring camera system and a robot vision. It is important not to change execution time in human detection system regardless number of persons. In this paper, the feature amount of an image was calculated by using Cubic Higher-Order Local Auto Correlation (CHLAC) in video sequence. CHLAC has position invariability and additivity. The proposed method can recognize a person and the other incoming things. We verify the human detection rate and the execution time by the number of persons. We show the system do not change the execution time in multiple human detectors.