The Theory And Techniques On Document Image Processing, Compression, And Decompression

Scanned image becomes the main fundamental representation of information in the office documents, and books etc. So, in the thesis, the author presents a document image processing scheme model, and mainly studies several key techniques in its implementation. The main works and conclusions in this paper are as follows:First, the paper states some features of document image, and the functional properties of a document image processing system. Accordingly, an architecture model of software in document image processing and compression is proposed.Second, on the basis of image processing principle, the author, who aims at the features of document image, offers a series of image processing techniques to compensate image quality and to wipe the noise. (1) Some specific technologies are discussed to solve the problems that lower the qualities of scanned image, so a lot of arithmetic operators are also given to sharpen the image. In this paper, there are some algorithms for the adjustment of image colors, brightness, contrast, and saturation. In the mean time, the image transformations, including blend, zoom, rotation, changing into gray-scale, mathematic morphology, etc, are supported. The thesis also gives some special functions to enhance the image effect, for instance, blur, mosaic, spray, mono-noise and color noise, etc. (2) A good many image filters are given to improve the visual effect, and form the base for the compression rate increase. (3) A theory for image smoothing and its application based on the correlation in quadratic distance is studied. (4) The author proposes an algorithm to extract the words from document image, and to exchange the image background, so that the noises can be filtered. (5) The author provides the support to export the result image as an image file in standard format, for instance, JPEG, GIF, and BMP etc. (5) Several quick image processing algorithms are offered, which are published in a Chinese key journal.Third, the author contrives the scheme for compression and decompression,which achieves a high compression ratio. The scheme aims at the specificities of document image, and the compression ratio can be changed dynamically. The distortion of decompressed image is low. Fourth, through the postprocessing according to the compress scheme, the effect of output image is reinforced.Fifth, in order to popularize the achievement, the paper also recommends a framework of extending application, and the open API is designed. The author developed a distributedcomponent to supply open interface, which accommodates to all sorts of applicable situation and secondary development.Finally, all stated above are carried out in the demonstration software .It can be concluded that the document-image processing model combines the theories and the practices. The author thinks that some ideas, methods, and implementation techniques discussed in this thesis are directly applicable to the developers during their process of developing digital library and virtual office based on network.