SPIHT's Improvement And The Application Of Mathematical Expectation And Standard Deviation In Static Image Compression

With the development of information technology, the image compression is alwaysa hot spot which is studied. How to reduce the complexity of compression algorithms,how to determine quantification threshold of transform coefficients in static imagecompression, how to choose wavelet bases in the image wavelet compression, and soon are still worth studying. According to the domestic and foreign researches, theauthor has done deep study and put forward a series of algorithms to solve abovequestions, the main contents are as follows:1. Have done the research of EZW and SPIHT algorithms, proposed improvementSPIHT image coding algorithm. Not only has the improvement algorithm inheritedoriginal SPIHT algorithm's merits, but it also reduced the SPIHT algorithm'scomplexity, and enhanced the original algorithm's working efficiency. The author hasimproved the original SPIHT algorithm through adjusting sorting process of waveletcoefficients, substituting for the original chain table structure with one dimensionalarray, changing the judgment basis of important wavelet coefficients.2. Have done the research of JPEG standard flow, applied the mathematicalexpectation and standard deviation in the image compression based on DCT. Using themathematical expectation and standard deviation to calculate the quantificationformula of DCT transformation coefficients, the different child blocks used differentquantification steps. Compared with using the fixed quantification table, the newmethod can be better reflecting the distributed characteristics of transformationcoefficients, and has good image compression effect. But this method only is suitableto the image which contains sole information.3. Have done the research of JPEG2000 standard flow, applied the mathematicalexpectation and standard deviation in the image compression based on DWT, andcarried on the wavelet coefficients' quantification. According to the mathematical expectation and standard deviation of each high frequency coefficient matrix, theauthor has constructed different quantification formulae with the positive and negativecoefficients. The experiment results show that the image compression effect is betterand the compression ratio is higher.4. Have done the research of the wavelet base regularity, determined the basis ofchoosing wavelet base in the wavelet compression. In the mathematical expectationand standard deviation wavelet compression, regarding images which contain soleinformation: the lower the regularity is, the higher image compression ratio is, andthe better the restructuring effect is. Regarding images which contain rich information:the higher the regularity is, the better the compression image restructuring effect is.