Spatially adaptive threshold for image denoisng based on nonsubsampled contourlet transform

In recent years, the threshold for removing noise based on wavelet transform has been very widely used because of its effectiveness and simplicity. Thus, there has been threshold based on a variety of frequency-domain transform. During the process of denoising, due to the differentiation of transform coefficients generated by noise and edge information, a good threshold for denoising can make a significant impact on the image quality. In currently existing threshold, spatially adaptive threshold based on Context-Modeling is proposed because of having considered neighboring coefficients so that it can adjust to coefficient characteristics. In this paper the improved spatially adaptive threshold method is applied to the nonsubsampled contourlet transform. Experimental results show that the method yields superior image quality and higher PSNR.