Medical Image Enhancement Algorithm Based on Improved Contourlet
Guo Qi,Shen Shu-Ting,Ren Ping-Chuan
DOI: https://doi.org/10.1166/jmihi.2017.2123
2017-01-01
Journal of Medical Imaging and Health Informatics
Abstract:To enhance the medical image, this article applies contourlet that exists beyond the wavelet theory, and most of the existing algorithms can not resolve the insufficiency of contourlet "does not possess the translation invariance, after the singular points of the image through the enhancement processing, artifacts still appear" which have been improved. First, the threshold value and threshold value function is improved and the multiple translation gains mean method is used to reprocess the denoising result, improving the "illegal Gibbs" effect. Second, we combine the respective characteristics of the low and high frequency subgraphs, enhancing two different proposed gain functions respectively, the enhanced image noise is effectively suppressed and the texture is clearer. We enhance mammogram in the simulation experiment by comparing it with many existing methods, such as the histogram method, Laplace method, improved wavelet method and traditional contourlet method. The enhancement effect of the suggested method is the best, image brightness and overall contrast are improved, the visual effect of the overall image is better, the detail section is enhanced, the outline of calcification is observed clearly, and white shadows cannot appear, and breast texture is also very clear.