EFFICIENT IMAGE CODING OF PYRAMIDAL STRUCTURES FOR THE HUMAN VISUAL SYSTEM

该研究项目研究了由小波核生成的图像的多分辨率金字塔分解，这种核与人类视觉系统处理兼容，并且提供了有效的图像编码。分析了与给定比特率的最佳视觉效果相关的生成核及其基础的截断、熵约简和率失真理论性质。静态和运动(视频)图像都会被处理。

This research project investigates multi-resolution pyramid decompositions for images, generated by wavelet-type kernels, which are compatible with human visual system processing and which provide efficient image coding. The generating kernels and their bases are analyzed for their truncation, entropy reduction, and rate-distortion theoretic properties relevant to the best visual result for a given bit rate. Both still and moving (video) images are treated.