UNCLOUD-BYKYLELI: UNIVERSAL IMAGE DEFOGGING USING IMAGE INPAINTING TECHNIQUES

Li, Kyle

Permalink

https://hdl.handle.net/2142/114933 Copy

Description

Title

UNCLOUD-BYKYLELI: UNIVERSAL IMAGE DEFOGGING USING IMAGE INPAINTING TECHNIQUES

Author(s)

Li, Kyle

Issue Date

2022-05

Keyword(s)

uncloud-bykyleli; image defogging;

Abstract

In recent years, image defogging has made significant progress. However, many defogging methods either require expensive sensors, or lack the ability to achieve optimal defogging results under all types of foggy conditions using a single model. In this paper, I propose uncloud-bykyleli, an image-based universal image defogging method. uncloud-bykyleli utilizes an unconventional approach that treats foggy areas of an image as irregular partial-holes, and attempts to recover a non-foggy image by using image inpainting techniques on the foggy image. Experiments on benchmark images demonstrate that in comparison to other singular image-based defogging models, uncloud-bykyleli achieves superior performance under all types of foggy-weather conditions.

Type of Resource

text

Language

en

Handle URL

https://hdl.handle.net/2142/114933

Owning Collections