Research On Adaptive Image Coding And Decoding

With the fast development of the computer technology and the Internet, multimedia application has grown in scope and stature. Simple text and voice can not meet people’s needs in daily communication anymore, profound changes have taken place and will continue to impact the education, production, lifestyle and daily work. However, in the meantime, multimedia data which are mainly composed of image data become increasingly large, which raises the pressure of storage media capacity, channel bandwidth and computer processing speed. Though by making the storage media with larger capacity, widening the channel width, improving the computer processing speed can relieve the pressure, but we will pay a heavy cost when we try to achieve these goals in reality. So the best way to solve the problem is by compressing the image before storing or transferring.So far, due to the excellent compression performance and low computational and memory complexity, JPEG becomes an attractive method for image compression. The most popular lossy image compression method used on the Internet and in the digital cameras is JPEG standard. However, when high quantization parameter is used to obtain a high compression ratio, JPEG images always suffer from disturbing blocking effects. Even though previous research has shown that a modified JPEG image codec by downsampling prior to encoding and upsampling after decoding can make the situation better, but when compression ratio becomes low, information loss caused by image downsampling is worse than JPEG quantization, compression performance of the modified JPEG image codec is not so good as that of JPEG, also, bit rate at the intersection of the rate-distortion (R-D) curves of two image coding algorithms is image dependent, thus under a given target coding bit rate, one can not tell which coding method is of better compression performance.Based on the research of image interpolation methods, this thesis proposes a space-variant cubic spline interpolation algorithm. And then, with the help of bit rate estimation method, adaptive image codec based on cubic spline interpolation and adaptive image codec based on space-variant cubic spline are proposed. By estimating the coding bit rate, the proposed algorithm chooses the image coding method with better compression performance from modified JPEG codec and JPEG standard. The experimental results show that the proposed algorithm can not only improve the R-D performance at low bit rates but also maintain the same performance with JPEG standard at high bit rates.