Because of existing shortcomings in wavelet denoising for the hard-threshold and soft-threshold method

a new multi-scale adaptive algorithm on threshold selection is presented.According to the characteristics of wavelet transform and 3σ-rule in noise signal

a self-noise reduction factor is multiplied by different wavelet coefficients.A new algorithm assessment of image denoising based on matching degree and image smoothness is proposed on the existing assessments of denoising algorithms.Experimental results show that multi-scale adaptive algorithm on threshold selection can effectively remove the image noise

retaining the image details.It significantly superiors to the traditional soft and hard threshold algorithms.