DSpace logo

Please use this identifier to cite or link to this item: http://dspace.bits-pilani.ac.in:8080/jspui/handle/123456789/16060
Title: An Alternate Approach for Single Image Haze Removal Using Path Prediction
Authors: Rajput, Amitesh Singh
Keywords: Computer Science
DehazeNet
Foggy conditions
Issue Date: May-2023
Publisher: Springer
Abstract: Haze removal is an important problem that existing studies have addressed from slight to extreme levels. It finds wide application in landscape photography where the haze causes low contrast and saturation, but it can also be used to improve images taken during rainy and foggy conditions. In this paper, considering the importance of haze removal and possible limitations of a well-known existing method, DehazeNet, we propose an alternate end-to-end method for single image haze removal using simple yet efficient image processing techniques. DehazeNet is among well performing haze removal schemes in the literature, however the problem of coloration and artifacts being produced in the output images has been observed for a certain set of images. Addressing this problem, the proposed solution is devised by taking advantage of the color features of the three color channels of an image to establish a decision criteria. Based on that, a suitable dehazing method for an input hazy image is selected. Removing the problem of poor coloration in output images, we have come up with an alternate method to remove the haze while retaining the visual and perceived quality of the image. The experimental results show that the proposed method yields better structural restoration, reduces haze content significantly, and does not cause any artifacts in the output image.
URI: https://link.springer.com/chapter/10.1007/978-3-031-31407-0_26
http://dspace.bits-pilani.ac.in:8080/jspui/handle/123456789/16060
Appears in Collections:Department of Computer Science and Information Systems

Files in This Item:
There are no files associated with this item.


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.