VIDEO PROCESSOR AND VIDEO PROCESSING METHOD

摘要

PROBLEM TO BE SOLVED: To provide a video processor and a video processing method for accurately separating only a person from moving objects, and for accurately separating even a person who does not move by matching a face detection technology with a human body area extraction technology using a depth video.

SOLUTION: The video processor includes: a face detection part for detecting the face of a person from an input video; a background model generation/update part for generating a background model by using the depth video of an input frame and the detection result of the face detection part; a candidate area extraction part for comparing the depth video of a frame following the input frame with the background model to generate a candidate area as a human body area, and for removing the area determined to be a moving object which is not a person from the candidate area by using the detection result of the face detection part, and for extracting the final candidate area; and a human body area extraction part for extracting a human body area from the candidate area extracted from the depth video of the current frame by using the human body area extracted from the depth video of the previous frame.

COPYRIGHT: (C)2013,JPO&INPIT