ENCODER AND ENCODING METHOD

摘要

PROBLEM TO BE SOLVED: To contrive to solve a problem in which a user's face region is occasionally erroneously detected in a chat using a video.

SOLUTION: A face region detector 12 detects a face region in a frame of a time-varying image. A history of the detected face region is recorded in frame unit in a face region history memory 44. A tracking portion 42 refers to a detection history of the face region in frame unit recorded in the face region history memory 44, and then tracks the face region detected by the face region detector 12 throughout a plurality of consecutive frames, thereby correcting a detection result of the face region by the face region detector 12. An ROI determination port 46 determines an attended region under a predetermined standard based on the face region corrected by the tracker 42. A video encoder encodes the attended region at a different image quality from the other regions and forms an encode stream of the time-varying image.

COPYRIGHT: (C)2009,JPO&INPIT