ABSTRACT
The images in creating, transmitting, and decoding processes generally get distorted by different types of noise. It becomes very necessary to reduce or remove the noise from the image in order to get its actual contents. Noise reduction has become a required step for any sophisticated algorithm in image processing. It is an open problem that has received considerable attention in literature for several decades. To make a balance between denoising and blurring and obtain clean images is a challenging issue in image processing. Though this issue has existed for a long time, there is no completely satisfactory solution. Over the last two decades, the wavelet-based denoising methods have been applied to the problem of noise reduction, and they have been shown to outperform the traditional nonwavelet-based denoising methods with or without intelligence approach. This chapter presents a case study on various image denoising methods by considering intelligence approach, that is, fuzzy logic. The denoising methods may be classified as nonwavelet and wavelet methods. In nonwavelet methods, the pixel values are modified in their original form, whereas in wavelet methods, the pixel values are first transformed into wavelet (time and frequency) domain, and then these values are modified in order to reduce the noise. In a denoising method, the pixels (or wavelet coefficients) are processed with or without using the information from its neighboring pixels (or neighboring wavelet coefficients), which are called term-by-term and block-by-block approaches, respectively. The state-of-the-art denoising methods assume that the images are corrupted with the additive white Gaussian noise (AWGN). A quantitative comparison between the denoising methods for both objective and subjective qualities of the images is considered as these two parameters are widely used for statistical computations in terms of peak signal-to-noise ratio and structural similarity index measure. The experimental results demonstrate that (i) the wavelet-based denoising methods using block-by-block approach give better results than the term-by-term ones for all noise levels; (ii) the wavelet-based denoising methods using intelligence approach, especially fuzzy logic, give better results than the nonwavelet without using fuzzy logic; and (iii) the wavelet-based denoising methods using fuzzy logic give better results for higher noise levels than the nonwavelet with fuzzy logic in terms of visual quality of the image.
