Categorisation of random images into fog and blur based on the statistical analysis. (27th April 2023)

• Record Type: Journal Article
• Authors: Verma, Monika; Kaushik, Vandana Dixit; Pathak, Vinay
• Is Part Of: International journal of advanced intelligence paradigms. Volume 25:Number 1/2(2023)
• Page Start: 84
• Page End: 106
• Publication Date: 2023-04-27
• Subjects: statistical analysis -- categorisation -- point spread function -- cumulative probability of blur detection -- eccentricity -- textured segments -- fog -- blur -- cumulative threshold -- feature-based classifier -- homogeneous images -- heterogeneous images -- out of focus blur -- motion blur -- Gaussian blur images; Artificial intelligence -- Periodicals; Machine theory -- Periodicals; Fuzzy logic -- Periodicals; 006.305
• Languages: English
• ISSNs: 1755-0386

Abstract

Noisy images are a bottleneck to solve the image processing problems. The present paper aims to classify images as different types of foggy and blurry images. A feature based classifier called FB classifier has been proposed. Given an image the classifier is able to tell whether the image is clear or unclear, which type of distortion is there, either foggy or blurry and also the categories of different types of blur and fog. The quality of the images taken through any equipment depends on few factors: 1) medium in which the photograph is taken; 2) the movements of either the camera or the object or movement of both; 3) the quality of the equipment that is used for capturing. All the algorithms of classification or the removal of distortions are made to handle the above three scenarios. The three factors encompass all types of foggy or the blurry images. The images viewed are given different threshold values according to their properties and finally the cumulative threshold value decides which type of the image is it. The algorithm is simple to implement yet it is comparable to the state of art methods.