An Edge-preserving Adaptive Image Denoising using Discrete Wavelet Transform

Abstract

Captured images got corrupted due to noise in additive or multiplicative form. It is necessary to reduce these noises for further image processing while preserving the edges present in the image. In this work, an edge-preserving adaptive scheme for image denoising is proposed. The noisy image is decomposed using discrete wavelet transform. Then two thresholds are calculated by using the Bayesian estimator. These thresholds are used to denoise the transformed image using soft thresholding. The first adaptive threshold is used for flatten region and second threshold is used for edges region as the noise has low visual perception on the edges. Experimental results show that this scheme achieves the state-of-the-art performance for image denoising. The results of these adaptive schemes are compared by Peak Signal-to-Noise Ratio (PSNR) and visual perception with existing denoising techniques.