Easy and Convincing Ear Modeling for Virtual Human

In this paper, we describe the human ear creation in a system for 3-D photorealistic face modeling from frontal and profile images taken by an uncalibrated handheld digital camera. The ear part is segmented in profile facial image with a boundary detection algorithm, and then mapped to the created face model. Our ear modeling procedure has several advantages. (1) A highly automatic detection algorithm is proposed to locate profile ear boundary accurately and robustly despite of a variety of illumination, hair shape, gender and race. (2) Deliberately designed shape deformation algorithm is proposed to stitch the ear and head smoothly. (3) Synthesized texture is utilized and merged with input images to compensate occluded ear-nearby areas and create realistic results. Experimental results show that the created ear is smooth and convincing, improving the face appearance dramatically.

• Type: Conference
• Year: 2004
• Where: PCM
• Authors: Hui Zhang, In Kyu Park
• Keywords: Detection Algorithm, Ear Modeling Procedure, Human Ear Creation, Multimedia, PCM 2004