Enhancement and Segmentation of Low-Light Images Using Illumination Map Estimation based Level Set (IME-LS) Method

Abstract

One of the most difficult aspects of image segmentation is that it cannot be successfully segmented if the image is dark or degraded. In this paper, proposes a method for the segmentation of low-quality and degraded images is put forth which is called Illumination Map Estimation based Level Set (IME-LS) Method. This proposed model has been classified into two parts: Firstly, we use an enhancement approach using illumination map approximation to enhance the input image. In this approach, Illumination map is constructed then refined. The refined illumination map undergoes an Augmented Lagrangian algorithm and then we use a sped-up solver for a considerably more efficient outcome. Secondly, Then the segmentation procedure begins once the image is enhanced. The enhanced image is segmented using level set bias method through Fuzzy clustering. In this method we employ the Fuzzy C-Means (FCM) algorithm to categorize data points into clusters. The fuzzy C-Means algorithm separates different entities in the image based on their varied intensities and sorts them into various clusters. The level set bias approach then tracks the variational boundaries of the image. We have designed the integrated algorithm in such a way that the image is classified or grouped into various clusters using a novel fuzzy Clustering algorithm and the variational boundaries of those clusters are tracked by employing the level set algorithm. In this paper, we further perform quantitative and comparative analysis of the suggested technique with respect to other segmentation techniques to illustrate the efficiency and flexibility of the suggested model.