一种改进的DCT/小波包三维压缩方法

An Improved 3-D DCT/Wavelet Packet Compression

对一种结合运动补偿的离散余弦变换(DCT)小波包的三维压缩方法进行了改进。变换前,使用6参数仿射运动模型的全局运动估计和补偿,消除了由于摄像机运动引起的背景变化所造成的影响;场景探测结合了人眼视觉系统(HVS)的特性,更有效地克服了场景改变带来的不利影响。计算机模拟证明了这种方法的有效性,压缩比达到90∶1,平均信噪比达到33 79 dB。

There are respective advantages and disadvantages in any one of discrete cosine transform(DCT) and wavelet packet compression,so combining them effectively is a method that can exert respective advantages.In this paper,an improved 3-D DCT/wavelet packet compression algorithm using motion vector segmentation was proposed.The affine six-parameter motion model was utilized for global motion estimation to estimate the camera motion.Scene-change detection with human visual system(HVS) was proposed to get over unfavorable conditions of scene change.The results of applying this method to a typical sequence by computer simulation were shown.The estimated compression ratio is approximately 90∶1,and the average SNR is 33.79 dB.