FACIAL IMAGE MATCHING METHOD IN A MODEL-BASED ENCODING SYSTEM

摘要

PURPOSE: A method for matching a three-dimensional standard facial image with a texture facial image in a three-dimensional model-based encoding system is provided. CONSTITUTION: When a picture of a face is provided to a model-based encoding system, a texture facial image and a three-dimensional standard facial image are obtained therefrom. The size of the three-dimensional standard facial image is modified until its face profile matches the profile of the texture facial image for thereby generating a profile-matched standard facial image. Then, the position of feature elements of the texture facial image is determined from the picture of the face on the basis of a knowledge base feature extraction algorithm, and the position of feature elements of the profile-matched standard facial image is also detected. The position of the feature elements of the profile-matched standard facial image is changed based on the position of the feature elements of the texture facial image, whereby a feature-matched standard facial image is generated, and texture data of the texture facial image is projected on the feature-matched standard facial image.