Announcement for Downloading full text filePlease respect the Copyright Act.
All digital full text dissertation and theses from this website are authorized the copyright owners. These copyrighted full-text dissertation and theses can be only used for academic, research and non-commercial purposes. Users of this website can search, read, and print for personal usage. In respect of the Copyright Act of the Republic of China, please do not reproduce, distribute, change, or edit the content of these dissertations and theses without any permission. Please do not create any work based upon a pre-existing work by reproduction, Adaptation, Distribution or other means.
URN etd-0825106-180250 Statistics This thesis had been viewed 2714 times. Download 1182 times. Author Ping-hsuan Chung Author's Email Address No Public. Department Communication Engineering Year 2005 Semester 2 Degree Master Type of Document Master's Thesis Language English Page Count 106 Title PERCEPTUALLY OPTIMIZED RATE CONTROL FOR JPEG2000 CODING OF COLOR IMAGES Keyword perceptible distortion rate - distortion control JPEG2000 image compression standard JPEG2000 image compression standard rate - distortion control perceptible distortion Abstract JPEG2000 is a new image compression standard that has been proven superior to the conventional DCT-based JPEG coding standard in both objective and subjective performance. However, the conventional rate control scheme, known as post-compression rate-distortion (PCRD) optimization, is not efficient and thus hindering the JPEG2000 coder from cost-effective implementation. As realizing the PCRD algorithm, intensive computation is required to encode the whole image completely and generate the entire bit stream, for holding which memory buffers of appropriate size must be equipped even though a large portion of the encoded data may not be included in the final code stream. Most of the computation and memory usage become redundant especially when images are encoded at lower bit rates. In addition to this inefficiency, the distortion metric adopted in PCRD optimization is usually mean squared error (MSE) which does not reflect real visual quality due to that the measured distortion may not be perceivable. Therefore, there still exists considerable room for the improvement in coding efficiency by adopting perceptible distortion in rate-distortion control.
In this thesis, a three-stage rate control algorithm is proposed to reduce both the computational complexity and memory usage over the conventional PCRD method while maintaining visual quality. In the first stage, the number of coding bits allocated to each code block is estimated by its relative significance in terms of signal energy. The initial truncation point for each code block is then determined by the number of coding bits allocated, and at least one coding pass is accomplished for each code block. In the second stage, coding bits are allocated each time to the successive coding pass of the code block with the steepest rate-distortion slope until they are consumed. The truncation points obtained in this stage are optimized in a sense that the allocated coding bits remove the maximum possible distortion from the remaining signals. In the third stage, the truncation of each coding block is globally optimized by applying the PCRD algorithm to the code passes encoded in the first two stages. The numbers of coding bits allocated in the first and second stage are reasonably set to be larger than the target bit number in hope that the set of optimum coding passes found by the original PCRD algorithm can be included in the set of coding passes obtained in the second stage. The coding complexity is largely reduced by taking a small fraction of overall coding passes in the PCRD optimization in the third stage and by determining the truncation points without calculating the distortion of each coding pass involved in the first stage.
Simulation results show that the same visual quality of the compressed color image can be achieved at lower bit rates by the perceptually optimized JPEG2000 coder with the reduction in computational complexity and memory usage by more than 50% in average as compared to the original JPEG2000 coder and the JPEG2000 coder rate-distortion by the method of priority scanning with optimal truncation (PSOT).
Advisor Committee Chun-hsien Chou - advisor
Chung-lin Huang - co-chair
Files Date of Defense 2006-07-04 Date of Submission 2006-08-28