人脸序列图像的模型基编码

Model-Based Coding for Facial Sequence

模型基图像编码自 80年代初提出后 ,由于在低码率情况下有高的主观图像质量 ,作为视频编码的方法得到了广泛的重视 .文中对模型基编码方法作了回顾 ,从线框模型的建立、运动参数估计、几何失真评价、码率控制以及实时性能等开展了对人脸序列图像的模型基编码研究 ,提出了许多新的观点和实验结果 .

Since the model based coding being presented from early 80's, it has got extensive attention as a video coding method due to high subject image quality and low bit rates. Model based coding is reviewed in this paper. Many fields have been studied from multi wire frame building, motion parameters estimation, geometric distortion evaluation criteria, bit rate controlling, and realistic performance for facial sequence. Many novel viewpoints and results are suggested.