作者: Charles P. Pace
DOI:
关键词:
摘要: Systems and methods for processing video are provided. Video compression schemes provided to reduce the number of bits required store transmit digital media in conferencing or videoblogging applications. A photorealistic avatar representation a conference participant is created. The can be based on portions stream that depict participant. face detector used identify, track classify face. Object models including density, structure, deformation, appearance illumination created detected An object algorithm, which uses machine learning detection techniques, creates from parameters derived models.