Adaptive spatial and multiresolution approach for image denoising

The denoising of digital images degraded by various type of noises is a major problem in image processing. By implementing with different successive filtering procedures, a very good tradeoff between detail preservation and noise removal is obtained. Pixels that are detected as noisy are filtered, the others pixels remain unchanged. In this present paper, a different approach to the problem is proposed based upon the method of joint bilateral filtering and wavelet thresholding. By implementing various spatial and transform domain methods either individually or combinedly for image denoising on various noise added images and evaluate the performance of the same. The proposed method for image denoising is performed through a recursive application of simple steps and demonstrated on a number of standard images. The results obtained with the proposed method outperforms other state-of-the-art filters both visually and in terms of objective quality peak-signal-to-noise ratio (PSNR).