Fractal Compressed Image Coding Based On New Features And Discrete Cosine Transforms

Fractal image compression has the advantages of high decoding speed and high compression ratio.However,fractal image compression coding has the problem that encoding takes too much time.Under the premise of ensuring that the amount of image reconstruction is almost unaffected,shortening the time of image encoding becomes the focus of this paper.The specific work of this article is as follows:(1)Based on the feature vector method and the basic fractal algorithm,the image block is processed first,and the dimension reduction is performed by the principal component analysis(PCA)method,and the effective vector information in the image block is selected as the main component.Features to achieve dimension reduction analysis-the purpose of feature display.When searching for matching blocks,a domain search method is used,and the size of the search field is adjusted according to the threshold.A fast fractal image coding algorithm based on principal component features is proposed.The experimental results show that the algorithm can effectively improve the encoding speed and similarity without affecting the image quality.(2)Matrix filling can use some matrix elements to recover the original matrix effectively,which can effectively reduce the time of image encoding and decoding.This paper combines the matrix filling with fractal image compression based on discrete cosine transform to form an improved algorithm.The experimental results show that the algorithm can shorten the decoding time.Compared with the basic fractal compression algorithm,this algorithm can effectively improve the image quality.At the same time,the encoding time is shortened.(3)In order to further reduce the encoding time,the image block is first quadtree partitioned,and then the value domain block and the definition domain block are classified according to a certain principle.Based on the basic fractal algorithm and discrete cosine theory,this paper presents an improved quadtree partitioned DCT fractal compression coding.The simulation experiments show that compared with the traditional DCT-based fractal compression algorithm,the “block effect” problem is effectively solved,the image characteristics are obvious,the image search time is less,and the reconstruction image quality is also guaranteed.Fractals can often be used for the description of irregular graphics,making it one of the directions of the image compression coding method.