The Research Of Medical Image Enhancement Algorithm Based On Multi-scale Transform

With the increasingly development of science and technology, more and more accurate and innovation theoretical studies have been founded in image enhancement algorithm, geometric analysis based on multi-scale transform is much better than traditional image enhancement algorithm. The theoretical research used by the newly developed methods of multi-scale analysis is very similar, the wavelet transform、 contourlet transform and so on are all the application of this method. Contourlet transform has soon be applied to the field of digital image processing as one important tool of a multiscale geometric analysis, which embodies the importance and advance of the multi-scale geometric analysis for image processing.The research of image algorithm based on multi-scale transform is generally conducted between decomposing the original image and the synthesis of image. In fact, for the image enhancement with wavelet transform, firstly, do wavelet transform on the source image and decompose various frequencies into many subgraphs which have different resolutions, then, according to the characteristics of subgraphs with different resolutions in transform domain, different enhancement algorithms will be token, finally, synthesize each processed subfigure. Because of every subgraph with different resolutions strengthen specificly, the boundary or texture in coarse and the fine boundary or texture can be enhanced at the same time to get a very clear picture. But, in the study process of the higher dimension, people find that the geometric characteristics which the wavelet transform has are not fully used and wavelet transform is not a better representation of function, then on this basis, the method of image enhancement based on Contourlet transform is proposed. Contourlet transform analyzes the scale and direction separately, it is an image representation of multi-resolution and local area. Its transformation uses Laplacian Pyramid (LP) to obtain a multiscale decomposition, then, filter it using directional filter bank (DFB), and finally, according to the information of direction, make inverse transformation for it into the new image.In the content of this paper, firstly, introduce the traditional image enhancement methods, such as gray level transformation, histogram processing, image noise reduction and so on, secondly, analyze and do the research on the theory of wavelet transform, at the same time, compare the differences between the original image and the processed image by using the image processing tool—Matlab, finally, the superior theory of Contourlet transform is put forward. In this paper, the author compares the enhancement algorithms which use wavelet transform with those based on the Contourlet transform, both of these two algorithms use the medical X-ray images as testing digital image, and on the basis of its theory, according to the characteristics of medical image, the author does the deep study on the image enhancement algorithm specific for the medical image.