Study And Apply On The Medical Image Fusion Based On Wavelet Theory

Research pixel-level image fusion algorithm in this article, Focus on studying the fusionalgorithm based on wavelet and using in the medical image fusion.Firstly introduc the conceptof image fusion background and significance, and then comparison with experimental ofimage denoising, image registration, image fusion algorithm.Because image denoising and image registration are the premise of image fusion. Firstly,the use of improved adaptive median filter to process the image denoising.And then improvedfilter to be compared with the conventional median filter, Experiments indicate that thedenoising effect is obvious. Once more image registration accuracy seriously affect the resultsof image fusion, Therefore, this paper made a research on image registrationalgorithm.Maximum information algorithm combining wavelet algorithm for imageregistration. Experiments show that this method can effectively improve the speed of imageregistration and effectively overcome the shortcomings of the registration slow speed andpoor real-time performance.Research basic theory of wavelet transform，Read a lot of literature and researchedmethods that base on wavelet transform image fusion field，Especially researched imagefusion rules of wavelet transform.For high frequency sub-band of wavelet transform selectethe rule that base on the window of pixels within the window energy contribution rate. Forlow-frequency sub-band, The image is divided into the same size sub-image,And thencalculateing correlation coefficient and standard deviation of their sub-images determine thelow-frequency fusion image coefficients. Finally, in this paper, the fusion algorithm is appliedto medical image fusion, simulation results verify the validity of the proposed method.