Design And Implement Of Image Compression Algorithm Based On All Phase Sequency Filtering

In the modern society, more and more multimedia information have been widely applied in daily life. The massive data of images and videos is used widely. Due to the data size of image and video is very large, it requires vast storage and wide transmit channel. It will exist a lot of problem in storing and network transmitting if it does not be processed. According to the development of information technology, more and more motionless and motion image compressing technique are used. Research on efficient compression algorithm is of great significance. In this thesis, a new image compression technique based upon the all phase sequency filtering is proposed and used to improve JPEG image compression standard. Compared to the traditional DCT (Discrete Cosine Transform), it has many advantages.In this scheme, three new transform, the all phase cosine biorthogonal transform (APCBT), the all phase inverse cosine biorthogonal transform (APICBT) and the all phase Walsh biorthogonal transform (APWBT) are used as an alternative of the DCT separately, then quantize the coefficients with the same quantization value to convert these coefficients to ZIG-ZAG sequence, finally encode the sequence with entropy-coding procedure. After reconstruct the image in term of reverse procedure and compare with traditional JPEG from compression rate, PSNR (peak signal-to-noise ratio) and other aspects. This scheme implements this system with VC++ and testifies the APCBT, APICBT and APWBT-based JPEG are superior to DCT-based JPEG with higher compression rate and better-quality of reconstruction image through extensive experiments. This scheme also discusses vector quantization, and combines it with APICBT to process image data, finally analyzes its capability and proper situation it can be used.