Study On Region-Of-Interest Coding Based On JPEG2000

The new generation still image compression standard JPEG2000 developed by JPEG not only provides improved compression performance, but also many new functionalities, such as: lossy to loss-less compression, embedded code-stream, ROI coding, progressive transmission by resolution, image security, the robustness of network transmission, random access to code-stream, error tolerance. ROI coding is a new feature in JPEG2000, ROI-based image coding is quite important in many image applications, it allows the user to set the prority of ROI higher than background, so the information pertaining to ROI is coded and transmitted earlier than the information pertaining to the background, to improve the image quality relatively. This technology can solve the contradiction between the image quality and the compression ratio.ROI coding algorithm based on JPEG2000 is studied in this paper. First, the academic background and purpose and significance of the research are introduced, the research status of the home and abroad is summarized; Second, the coding modules of the JPEG2000 coder and the critical technology—wavelet transform are introduced in detail; Third, the common questions of wavelet transform in image processing field are studied, and the effect of image compression with different wavelet bases and the optimal decomposition level are given by experimental simulation; Finally, after analyzing the merits and drawbacks of the two standard ROI coding algorithms in JPEG2000, a new ROI coding method called mixed coding method is presented. Experimental results show that the new method has the advantages of the two standard algorithms, it doesn't need to code and transmit the shape information of ROI, and can adjust the compression quality of ROI and background conveniently.