A new hybrid image restoration method based on fusion of spatial and transform domain methods

Abstract

A novel adaptive spatial and transform based approach by fusing Bilateral Filter (BF), Joint Bilateral Filter (JBF) and Wavelet Thresholding (WT) is proposed for image restoration. The Bilateral and Joint Bilateral filter can perform as an edge-preserving smoothing operator and have better behavior near the edges. In the first stage, the noisy image is passed through bilateral filter and some amount of noise is reduced but the image becomes blurred, hence wavelet thresholding is applied first with Bayes Shrinkage (BS) rule and then with Modified Bayes Shrinkage (MBS) rule in the second stage. In the third stage, the wavelet thresholding output is used as a reference image for the joint bilateral filter to preserve and enhance the edges effectively. The main aim is to achieve a cleaner version of the noisy image without blurring important features like edges, curves, and textures. The filter-performances are usually compared in terms of peak-signal-to-noise ratio (PSNR) and visual quality sense. Experimental results show that the proposed image denoising method is competitive when compared to other methods in suppressing various types of noise.