Jae Hyun Park, Kyu Sik Park, Syng Yup Ohn, Chun Kuk Kim

The KIPS Transactions:PartB , Vol. 8, No. 1, pp. 20-27, Feb. 2001
10.3745/KIPSTB.2001.8.1.20,   PDF Download:

Abstract

Face detection is a very active research topic having many applications like automatic surveillance or intelligent recognition systems. In this paper, we are focusing on the development of an effective face detection algorithm, which becomes a basic tool for intelligent video conference systems. Here, intelligent system means that video camera should have the capability of both detecting and tracking the speaker's movement. The input images are transformed from the RGB to the YIQ color space. Then, the IQ components are used for skin region detection using the skin color table and facial features are searched through the Y components. Mask operation is also involved in detecting process to remove the falsely detected areas. In our approach, three facial features area, shape and luminance ratio, are utilized. From the experimental results, we confirmed that multiple faces in an image with a complex background can be detected fairly well through the proposed algorithm.

Statistics
Cite this article