Motion-Compensated Video Coding with Adaptive Perceptual Quantization

The Motion Picture Experts Group (MPEG) has standardized a syntax for the coded representation of video. The standard specifies only the bit-stream syntax and thereby a method of decoding. It allows much flexibility in designing the encoder for optimal trade-off of complexity versus performance. The syntax supports the basic encoder operations of motion compensation, quantization of discrete cosine transform (DCT) coefficients and variable-length coding. The operation key to reconstructed picture quality is that of quantizing the DCT coefficients. Adaptive quantization techniques conforming to the MPEG syntax can significantly improve the performance of the encoder. In this paper, we develop a strategy for adaptive quantization that uses a perceptual basis for optimizing picture quality. We concentrate on a one-pass causal scheme to limit the complexity of the encoder. The system employs prestored models for perceptual quality and bit rate that have been experimentally derived. A framework Is provided for determining these models as well as adopting them to locally varying scene characteristics.