A Comprehensive Study on Wavelet Based Shrinkage Methods for Denoising Natural Images

Transmitting the information in the form of images has drawn much importance in the modern age. The images are often corrupted by various types of noises during acquisition and transmission. Such images have to be cleaned before using in any applications. Image denoising is a thirst area in image processing for decades. Wavelet transform has been an efficient tool for image representation for decades because of its simplicity, energy compaction and sparse representation. Ample of wavelet based thresholding techniques are proposed based on universal and adaptive thresholding techniques. Fixing an optimal threshold is a key factor to determine the performance of denoising algorithms. This optimal threshold shall be estimated from the image statistics for ensuring better performance of noise removal in terms of clarity (or quality of the) images. In this paper, an experimental study of the state of the art wavelet based thresholding methods is presented. The denoising performance of the wavelet based shrinkage methods are compared interms of mean square error, peak signal to noise ratio, image enhancement factor and the most recent measure namely multiscale structural similarity index. Key-Words: Image denoising, Wavelet transform, Threshold methods, Adaptive threshold, Wavelet subbands, Shrinkage methods.