High-efficiency Image Coding

Image signal is used more frequently in multimedia signal processing.It has also been noticed that the high-definition and high-fidelity images are becoming more and more popular in practical applications.However,the bigger data-size of such images always limits the application of them.To solve this problem,people propose to compress the source image data before the storage and processing to get an effective reduction of the data-size.In image compression,the spatial redundancy of the source image is removed by using some efficient coding tools,which accordingly makes a reduction of the data-size.However,the compression of the source image signal always leads to the distortion.To reduce the distortion and improve the quality of the compressed image signal,the more efficient image coding algorithms should be carried out in practice.In this thesis,we focus on the study of high-efficiency image coding based on some advanced compression techniques.More details are presented in the following.(1)We propose an efficient image coding method by using the pixel domain padding algorithm and the transform domain downward conversion.(2)Based on the analysis of the compression distortion relationship between the pixel domain and transform domain,we develop an error compensation algorithm and implement a more efficient image coding through applying it to the transform domain downward conversion based coding.(3)We apply the interpolation-guided adaptive filtering to both of our proposed coding methods mentioned above and improve their coding performances more.Experimental results demonstrate that our proposed methods outperform the existing image coding methods and a significant performance improvement has been achieved.