Research And Implementationon On ROI Coding Based On JPEG2000

With the development of multimedia and network technology, it requires to storage and transport more images in the limited space and bandwidth, and providing different resolution or quality of the review picture according to actual needs. According to users'subjective concern, ROI coding divides an image into region of interested and background, and different coding makes ROI at a high priority, or in decoding, it can first restore the image of ROI and the quality of rebuilding better than the background. ROI coding could not only satisfy requirement for restoring important parts of the image at high quality and even lossless in the low bit rate, but also can be a good solution of storage space and bandwidth bottlenecks at process of communication, combining with image compression algorithm. This feature makes ROI coding becoming a hot research point in field of image coding. JPEG2000 is the latest static image compression standard, supported the lossless and lossy compression, besides basic image compression capabilities, it also has many features,and supporting ROI coding is the most important feature of these.The paper mainly researches ROI coding technology based on JPEG2000 standard, and presents two types of ROI coding methods combining with rate-distortion optimization and coefficients shift. Firstly, analysis the JPEG2000 coder process, the key technology and code parameters in detail. Then analysis and summarize ROI coding method based on coefficients shift. According to the data distribution feature after wavelet transform, present a new ROI coding algorithm which based on coefficients shift and used different shift factors in different subbands. The low frequency coefficients play an important role on image subjective quality, this algorithm more conform to Human Visual Properties. ROI coding methods based on rate-distortion optimization, changes the order of coding stream through advancing slope of a code block reserved intercept points. The methods based on coefficients shift increases bit planes always reduced coding efficiency, while the methods based on rate-distortion optimization is restricted to accuracy of ROI by the size of code block. JPEG2000 supports ROI coding methods based on coefficients shift and rate-distortion optimization, according to the data distribution feature after wavelet transform, presented a new slope promotion strategy based on subbands, and a new algorithm integrate with Maxshift method. Slope promotion only used in the coding, without any steps in the decoding. The experimental results show that the new algorithms can be adjusted the contrast of quality of ROI and BG, while it has the same complexity, coding efficiency and accuracy of ROI in maxshift algorithm. For achieving different priorities in multiple roi coding, combined with slope promotion strategy and PSBShift algorithm, PSBShift shift factor sets the priority of ROI,and slope promotion adjusted the parts of quality of reconstructed image. Simulation results show that the algorithm is applied to good performance both single roi and multiple roi. We explore preliminary nonlinear function in order to choosing the suitable stategy and understanded relationship between order of coding stream and slope promotion. In most cases, slope promotion strategy based on subbands can be adjusted the contrast of quality of ROI and BG, then the reconstructed image conform to Human Visual Properties.