JPG文件结构分析(国外英语资料).doc

JPG文件结构分析(国外英语资料) JPEG is a compression standard, and can be divided into standard JPEG, progressive JPEG and JPEG2000 three: Standard JPEG: storing a single raster image in 24 bit color is a platform independent format that supports the highest level of compression, but this compression is lossy. This type of picture can only be displayed from top to bottom when the page is downloaded, but the picture can not be seen until all the pictures have been downloaded. The gradual improvement: JPEG format, progressive JPG standard JPG support staggered, the website can be downloaded, showing a picture of the first rough appearance, then slowly showing the complete contents of progressive JPG files is less than the standard JPG file to. JPEG2000: the new generation image compression method has better compression quality, its compression ratio is about 30% higher than the standard JPEG, and it also supports lossy and lossless compression. A very important feature is that it can achieve progressive transmission, that is, to transmit the outline of the image, and then gradually transfer data, so that the image from hazy to clear display. Taking a 24 bit color image as an example, the compression of JPEG is divided into four steps: Color conversion: before you can compress a color image, you must first convert the color pattern to data. Data sampling is also needed after the conversion is complete. DCT transform: transform the image signal in frequency domain, separate the high frequency and low frequency information, and then compress the high frequency part (namely image detail) of the image. Firstly, the image is divided into several 8 * 8 matrices, and then the DCT transform of each matrix is made. Turn the 8 * 8 pixel matrix into a 8 * 8 frequency coefficient matrix (the so-called frequency is the speed of color change), and the frequency coefficients are floating-point numbers. Quantization: since the codebook used in the following fourth steps is integer, it is necessary