The Study Of Medical Image Enhancement Based On Wavelet

Image enhancement is one of main contents of image processing, improving the intelligibility of the image, that is highlighting some features of the image selectively, make it more applicable to specific areas is the main objective of image enhancement. Breast cancer is one of tumor, the human tissue, illumination, noise and other factors make the calcifications and masses of the image are not clear, affecting the doctors'identification of lesions and normal tissue by X-ray images. For this, if enhance the X-ray images before the doctors'diagnose, prominent the calcifications and masses of the images, that we can reduce the diagnostic difficulties caused by the poor picture clarity effectively, to reduce the rate of misdiagnosis.There are many image enhancement algorithms based on wavelet, but these methods all focus on dealing with wavelet coefficients, they all decompose the low-frequency sub-image to get the details, we notice that there are great details in high-frequency sub-images, if further decompose the high-frequency sub-images, we can extract more details. Based on this, this article propose a new medical enhancement algorithm, it improve the traditional method of image decomposition, and with ant-symmetric biorthogonal wavelet to decompose the image, overcome the shortcomings of large computation and not satisfy the real-time when detect edges.The purpose of enhance the breast image is to highlight the calcifications and masses, wavelet image enhancement algorithm can well enhance the calcifications, but tumor characteristics need curvelet, which can detect curves to enhance. Basically, the existing methods do curvelet transform for each sub-images at the same time, as the curvelet is based on Radon transform, complex computation and easily bring blockings, based on this, this article improved the wavelet enhancement algorithm proposed with curvelet, detect edges by curvelet and refine edges by wavelet.Simulation experiments of mammography X-ray images are implemented by Matlab , the results show that the proposed algorithms have many advantages over other algorithms, such as small computation, the edges and textures of the image are clear after dealing, the calcifications are independent, no caking and large white.