Application Of Multi-scale Analysis In Medical Images Enhancement

With the coming of the information age, especially the digital age, the medical image has become an important assistant means for the diagnosis and therapy of the doctor. The development of medical image technology advances the request for the post-processing, while image enhancement, as an important technology of the image post-processing, improves the quality of medical images and the validity of diagnosis. And the practical value of medical images can be improved by using of multi-scale geometrical analysis, which utilizes the time-frequency features effectively, showing interesting information prominently, and restraining some hashes at the same time.This paper made some beneficial attempts to the image enhancement question, embarking from medical image multi-scale analysis. First, Wavelet transform combines with a nonlinear transform to de-noise and contrast enhance the CR images, realizing the retention of detail enhancement effect, filtering the noise simultaneously. Next, Wavelet combines with homomorphism filtering and GAG feature curve to de-noise and enhance edge in order to enhance contour of images obviously. Finally, using Contourlet transform based on invariance translation to de-noise, which obtains the even better effect than Wavelet transform. The experiment results using these methods are in good agreement with the expectation, which establish the foundation for further studies.