Study on adaptive thresholding technique of image denoising based on wavelet transform

Abstract

It is studied in the paper that an adaptive soft and hard thresholding image denoising method, in which image pyramid decomposing is realized by wavelet transform, and the mean value, mid-value and root mean square value of different sub bands are calculated as thresholding. The image is added into different kinds and different intensities noise, and processed by different wavelet decomposing levels and thresholding selected algorithms, the total 27 kinds of thresholding combination schemes are completed in the research process. The SNR (signal noise ratio) and PSNR (peak signal noise ratio) of denoised image are compared and analyzed and benefited results are achieved. Furthermore, the algorithm in reference is realized by MATLAB program, the results of reference are compared with that of the paper to demonstrate the significance and correctness of the results in the paper.