Wavelet Thresholding Algorithms for Image Denoising Aditya Rana 1 and Charu Pathak 2 1 Department of Computer Science and Engineering, Manav Rachna University Faridabad; aditya.rk.rana@gmail.com 2 Department of Electronics and Communication Engineering, MANAV Rachna University, Faridabad; drcharu2004@gmail.com Abstract This paper talks about the wavelet thresholding algorithm for image denoising. Any data, either in the form of signals, or images contains more noise than informations. To make sense out of it, it needs denoising. For that, this paper explains algorithm that makes active use of wavelet thresholding to achieve maximum denoising. For statistical analysis matlab software is used as it comes with wavelet thresholding application. This is then used to process standard lenna image to obtain haar wavelet transform for three levels of decomposition of image. On the contrary daubechies wavelet transform is also applied to the same sample image of lenna. Using Haar Wavelet for image compression has a little bifurcation in Retained Energy and Number of Zeros along x axis. On the other hand Daubechies Wavelet compression with global thresholding on decomposition level 4 for standard image of lenna yields different trend lines between Retained Energy and Number of Zeros. Its applications vastly covers all medias such as image, video, signals, etc. to achieve maximum in- formation. With advances in image denoising, space can be utilized more appropriately as user can be able to save space in his personal devices like mobile phones, laptops, etc. With this user can be able to use or access that free space in order to upload more data, or use it for his computational use. Keywords: Image Denoising, Thresholding, Wavelet Transform 1. Introduction Digital Images are ofen corrupted with noise during acquisition, transmission, and retrieval from storage media. Many dots can be spotted in a Photograph taken with a digital camera under low lighting conditions. Te denoising algorithm is used to remove such noise. Image- processing algorithms such as pattern recognition require a clean image to work efectively. Random and uncorre- lated noise samples are not compressible which leads to the importance of denoising in image and video processing. If Y is the observed noisy image, X is the original image and N is the AWGN noise with variance σ2. Te objective is to estimate X given Y. A best estimate can be written as the conditional mean X = E[X | Y]. Te difculty lies in determining the probability density function ρ(x | y). Te purpose of an image-denoising algorithm is to fnd a best estimate of X. Tough many denoising algorithms have been published, there will always be scope for improvement and betterment. Image noise is an important aspect in terms of processing the image. It ranges from sharp specks on a digital photograph taken in good light to optical and radio astronomical images that are almost entirely noisy, from which we can derive a small bit of information by sophisticated processing. Such high noise level in image will be unacceptable in a photograph since it would be impossible to even determine the subject in the image. An another model of image noise is Gaussian which is additive in nature, independent at each pixel, and independent of the signal intensity, caused primarily by Johnson-Nyquist noise (thermal noise). As it comes from the reset noise of capacitors (kTC noise). In color cameras, which have more amplifcation in the blue channel rather *Author for correspondence Indian Journal of Science and Technology, Vol 11(27), DOI: 10.17485/ijst/2018/v11i27/130706, July 2018 ISSN (Print): 0974-6846 ISSN (Online): 0974-5645