Research On Image For Sparse Decomposition Based On MP With FFT

Currently, digital image processing has been widely used in many different areas, such as scientific research, military, public security, education, economic construction and the various aspects of daily life. However, the digital image data is huge, so it is necessary to compress the image data. With the continuous improvement of the social informatization degree, the application of image information will be more widely, therefore further research on the image compression is of great significance.In recent years, a non-orthogonal decomposition method-image sparse decomposition is developed, and the result of decomposition (which we refer to image sparse representation) is very simple. Just this simple feature makes it an effective way for a low bit rate image compression. The image compression based on the sparse decomposition has been the most potential possible solutions for a low bit rate image compression. In many signal and image sparse decomposition methods, the Matching Pursuit (MP, matching pursuit) algorithm is proposed earlier. The MP algorithm is simple in principle, easy to understand and realize, hence it is currently the most widely used sparse decomposition method. But meanwhile this algorithm has a bottleneck, that is the decomposition complexity is high and the amount of calculation is very large. In order to resolve this problem, improving the speed of the image compression algorithm with a good compression, this paper uses the image sparse decomposition based on MP algorithm, converts very time-consuming inner product calculations in sparse decomposition into crosscorrelation calculations, and implements crossconelation calculations by fast Fourier transform.On this basis, in order to further improve the speed of operation, this paper uses VC++and Matlab mixed programming. With the powerful scientific computing functions and rich toolbox of Matlab, and the high execution efficiency of VC++, it can further enhance the algorithm efficiency. Finally, the experimental results show that the design proposed in this paper can improve the rate of decomposition of sparse effectively.