Image Denoising Algorithm Based on Structure and Texture Part
Yifeng Cheng, Zengli Liu · 2016
With people's pursuit of high quality image, image denoising has always been a popular research. The traditional image denoising method is based on wavelet transform threshold. The denoising effect is good, but is prone to lose the image structure and texture information. Based on the deficiency of the traditional denoising method, this paper put forward a method that is based on morphological component analysis (MCA) to decomposes an image into texture and structure. The part of texture uses all phase biorthogonal transform(APBT) dictionary sparse representation to denoise. The part of the structure uses Block-Matching and 3-D Filtering (BM3D) algorithm to denoise. Finally, combined with the two parts to get the final denoising image. Experimental result show that compared with the traditional wavelet threshold denoising, this paper's algorithm can better retain the image details and structure information, it can get better denoising performance in PSNR.