Digital Image Compression Based On 2-D Empirical Mode Decomposition

Digital image compression is a hot research topic in the field of image processing. The main problem of image compression is to minimize the data which express the digital image, and eliminate redundant data.This article carries on the brief introduction about image compression coding theory and the trend of development of it. Based on the analysis of traditional image compression methods and the use of two-dimensional EMD method in image compression, we carry on the analysis comparison for the image compression results. The main research works are as follows:(1) The empirical mode decomposition. This method has the advantage of the non-linear, non-stationary signal processing. We extend EMD method to two-dimensional and apply it to the field of image procession. Uses different interpolation methods and improve the algorithm on digital image, this paper find a suitable method to get the image signal envelope, completes the screening process;(2) Use the method of BEMD to carry on the image compression. BEMD method has provided a powerful tool for the image processing. After decomposed by the BEMD method, the image will be a series of IMFs and residue. As the first layer of IMF contains the details part of the image, it will cause some loss of high frequency part by only use one method to all IMFs and residue. Also it will affect the quality of image compression. The first layer of IMF will be treated differently with the others;(3) Compare with the results of the DCT compression algorithm, the extreme point compression algorithm and the improved multi-sample block overlapping compression algorithm, this paper proposes a new digital image compression method based on BEMD decomposition.In the above of this theory work, this paper completes the experiment simulation on the basis of MATLAB 7.0 environment, and carries on the analysis and summary to experimental results. The BEMD method which applies in image compression, saves the transmit channel in a certain extent, and agrees with the progressive display encoding requirement greatly. Through the comparison of experimental images and data, find out a better method for digital image compression.