作者: Pei-Ying Chiang , Tsai-Yen Li , Wen-Hung Liao
DOI:
关键词:
摘要: We developed a system for automatic generation of caricatures by analyzing unique facial features the subject. The proposed face model contains 119 parameters derived from MPEG-4 standard. These nodes are categorized into 8 groups with pre-defined hierarchy to guide placement components after shape exaggeration. Given an input image, determines which and how component should be altered. Using artist’s finished work as source is capable producing similar style effectively efficiently.