Research On Image Compression Algorithms Based On Non-symmetry And Anti-packing Model

With the advent of the information age,a huge amount of image information will be produced all the time.Often the quality of the images is not good enough due to limitations in the shooting equipment,environment or storage.Compared with the image we want,there are some problems such as blurring and distortion,which cannot be used directly.Therefore,in order to solve these problems,a series of image processing technologies emerge,such as image compression,image segmentation,image recovery and so on.Image compression is an important field of image processing.Its basic principle is to express images with less data,so that the total amount of data required for storage is reduced,so as to reduce storage costs and data transmission costs.The main work of this paper is to study the application of non-symmetry and anti-packing model combined with other algorithms in the field of image compression and image representation,and two algorithms are proposed:The first algorithm is an improved block truncation coding based on rectangle nonsymmetry and anti-packing model.The paper demonstrates the combination of the nonsymmetry and anti-packing model's block segmentation algorithm and block truncation compression algorithm is feasibility and superiority.The theory is verified by experiments and compared with several compression algorithms of block truncation.The experimental results show that the improved algorithm has better compression speed,compression rate and decoding image quality.The second algorithm is an image representation algorithm based on the overlapping rectangular non-symmetry and anti-packing model and binary bit-plane decomposition,which can be subdivided into two algorithms: the gray image representation based on the overlapping rectangular non-symmetry and anti-packing model and binary bit-plane decomposition and the color image representation based on the overlapping rectangular non-symmetry and antipacking model and binary bit-plane decomposition.The idea of the gray image representation algorithm is to decompose the gray image into several binary images and then compress each binary image by overlapping rectangular non-symmetry and anti-packing model.The proposed gray image representation algorithm is compared with the linear quadtree representation of gray image,the gray image representation based on triangle non-symmetry and anti-packing model and binary bit-plane decomposition and the image representation based on square nonsymmetry and anti-packing model and binary bit-plane decomposition.It can be seen from the experimental results that the number of segmentation blocks of the proposed gray image representation algorithm are greatly reduced,and the compression rate is higher than others.The proposed gray image representation can be used as a supplement to the gray image representation algorithm based on non-symmetry and anti-packing model and binary bit-plane decomposition.By the extension of the gray image representation algorithm,a color image representation based on the overlapping rectangular non-symmetry and anti-packing model and binary bit-plane decomposition is proposed.The idea of the color image representation is to decompose color image into three gray images based on different channels,and then process three gray images using the proposed gray image representation algorithm.Compared with the linear quadtree representation of color image and the color image representation algorithm based on rectangular non-symmetry and anti-packing model and binary bit-plane decomposition,the experimental results show that the proposed color image representation also has better compression ratio and fewer segments.