A New Image Denoising Method Based on Wavelet Multi-scale Registration Fusion
Liang Gao, Yue Ma, Shuai Chen, Jing Hui Wu · 2018
Image denoising is an eternal research topic. In this paper, a new image denoising method based on wavelet multiscale registration fusion is proposed to solve the problem that it is easy to lose the edge and texture details of the image in the denoising process. First of all, we can get multiple sets of wavelet coefficients by using different wavelet bases to decompose the same noisy image. Then, the obtained wavelet coefficients are processed by the improved wavelet threshold shrink to get multiple denoising images of the same noisy image. At last, we use the fusion registration algorithm proposed in this paper to fuse the edge feature of multiple denoising images to get the final denoising image. The experiments prove that this method not only can effectively overcome the pseudo gibbs phenomenon caused by the hard threshold method, but also can overcome the image distortion phenomenon caused by the soft threshold method. More importantly, compared with existing methods, this method can effectively preserve the edge detail and texture features of the image and the image has a better visual effect after fusion registration. Therefore, it has a better application value.