A Unique Approach To Blind Image Deconvolution

Abstract

One of the primary objective in digital image processing is extracting features or displaying useful information. It is an active area of research in the modern age of computing. Image acquisition is the very first step in almost all application areas of digital image processing. Due to limitation of image sensing hardware some visual artifacts might be introduced in the image. Thus, after acquiring an image to use it in applications the acquired images are required to be enhanced. Sometimes, to obtain a high quality image the low level processing is required. Wherein image data itself has to be manipulated to reconstruct a good quality image. Image enhancement is subjective whereas restoration is an objective process. Restoration of an image requires detection of the degrading function. Then applying inverse of such degrading function can recover the image. It needs a priori knowledge of degrading function. Image deconvolution is one such technique of image restoration. When the deconvolution function is not known in advance, it is called blind deconvolution. This paper shows a unique approach involving image restoration and image enhancement techniques to reconstruct the clear images from the degraded images. The quality of a randomly selected image can be restored using blind deconvolution method. Image enhancement techniques can be used to highlight certain features of interest as well as to confine few details. The proposed method decreases image noise, annihilate artifacts and produce clear view of details of the image. It amplifies certain features of image to be displayed. Finally, an acceptable version of images closer to the true image is obtained. The purpose of research is to enhance the features of image using nonlinear techniques along with blind image deconvolution. Our proposal is suitable for noisy, blurred images. It can be generalized to the images that have been degraded by other means.