Wavelet-DCT Based Image Coder for Video Coding Applications

摘要

Discrete Cosine Transform (DCT) is widely used in intraframe compression to compress the current frame of a video source. With little differences between frames, interframe compression uses the subsequent frame to compress the reference frame, by performing subtraction. The subtraction output (difference image) is fed into the intraframe coder for still-image compression. However, DCT cannot compress difference image efficiently, since it is Discontinuous Tone Image (DTI) [1] and the algorithm should be improved. Wavelet-DCT is a novel algorithm that uses Forward Discrete Wavelet Transform (DWT) to compute DCT. A SIMULINK model for an intraframe coder is developed and tested. Then, a Wavelet-DCT block is coded using MATLAB and used to replace the conventional DCT block. In the study, on one hand, experiment is conducted on difference image for conventional intraframe coder; on the other, the same difference image with Wavelet-DCT based intraframe coder. The results prove that Wavelet-DCT has greater compression for difference images. This is possible since the algorithm allows discarding insignificant DWT coefficients or more popularly known thresholding the DWT coefficients while computing the DCT. We then conclude that Wavelet-DCT is capable compressing difference images at higher compression in video coding applications.