Fractal Image Compression Coding Algorithm Based On Methods Of Feature-Vector

In the information age,huge data has become one of the main characteristics of today's society.The problem of how to store large amounts of data,especially images' data,has also received increasing attention from society.Therefore,image compression technology,which is one of the core technologies to facilitate data information storage,has become the main research direction of many scholars.At the same time,fractal image compression technology has become one of the most advantageous image compression technologies due to its high compression ratio.However,the traditional fractal image compression coding algorithm consumes a lot of time,which seriously hinders the practical application of fractal theory in the field of image compression.Aiming at this drawback,four new algorithms based on feature vector and fractal theory are proposed to improve the encoding speed and improve the quality of decoded images.The main research process is as follows:(1)Through the research on fractal algorithm and feature algorithm,a algorithm about sum of four-line and feature coding is proposed,which converts global search into local search(Near neighbor search)according to the relationship between matching root mean square error and sum of four-line's eigenvalues,which limits the search space and reduces the search for domain blocks.The simulation results show that the decoded image is better than the 1-norm feature algorithm in objective quality.Compared with the basic fractal coding algorithm,the four-line and feature algorithms do not change the subjective quality of the reconstructed image,but at the coding speed.Greatly improved.(2)For the problem that the objective quality of the decoded image of the sum of four-line's feature algorithm is not ideal,a sub-block mean point feature algorithm is proposed to reduce the lack of pixel information of the image block.The algorithm is compared with sum of five-point's feature algorithm,1-norm feature algorithm,European-style feature algorithm and double-crossing algorithm.The simulation results show that the algorithm in the paper is less objective and the objective quality is better.excellent.(3)Through the research on fractal algorithm and feature algorithm,a new algorithm is proposed in this paper.The algorithm uses the variance to reflect the information between different image blocks,and takes the variance ratio between the image block sub-block and the parent block to reflect the information of different parts of the image.The theoretical proof and experimental simulation show that the proposed algorithm is better than the basic fractal coding algorithm in coding speed,and better than the 1-norm feature algorithm in decoding image quality.(4)Through the research on fractal algorithm and feature algorithm,this paper proposes a unit European product fractal algorithm,which not only changes the full search into a local search,but also fully reflects the connection between the sub-block and the parent block.Simulation experiments show that the proposed algorithm is nearly 100 times faster than the traditional fractal coding algorithm,and the PSNR value of the algorithm is better than the 1-norm feature algorithm.