A Study on Bi-Level Image Compression Using Improved MMR Coding

• Main Authors: Feng-Fu Tseng, 曾豐富
• Other Authors: Ching-Huang Wei
• Format: Others
• Language: en_US
• Published: 2002
• Online Access: http://ndltd.ncl.edu.tw/handle/31757504962627555274

碩士 === 國立高雄第一科技大學 === 電腦與通訊工程所 === 90 === ABSTRACT In this thesis, we review several lossless compression algorithms for bi-level image compression to enhance the understanding of bi-level image coding and propose the improved version of modified MMR. MH (Modified Huffman) coding is a one-dimensional run-length coding (1-D RLC). MR (Modified READ) coding is a mixture of one- and two-dimensional (1-D and 2-D) coding, and MMR coding is only based on the 2-D MR algorithm. However, MR coding and MMR coding are line-to-line reference models. The compression codewords are based on the distance calculated between the next coding pixel and the reference pixel. Modified MMR coding inherits the codewords from MMR coding except the difference that Modified MMR coding uses macroblock-based reference instead of line-to-line reference. It is an effective binary shape coding for digital video compression. Improved MMR coding proposed by us modifies the procedure of modified MMR algorithm to reduce the complexity of the coding. MATLAB programming is utilized to implement the simulation for improved MMR coding and modified MMR coding, and the experimental results show that both of them could achieve lossless compression coding without any image distortion. Taking the compression rate (CR) to be compared between improved MMR and modified MMR coding, we find that the improved MMR coding obtains higher compression rate in order to reduce more storage space than modified MMR coding. And, the algorithm of improved MMR coding is easier to be implemented by using MATLAB programming, too.