Enhancing Underwater Imagery Clarity: A Novel Approach with Zero DCE NET

Underwater imaging poses different challenges to the enhancement techniques because of the distortion, quality, absence of natural light and many other factors. In this paper, we have explored the efficiency of deep learning based algorithms in the realm of image enhancement. Our methodology explore...

Full description

Saved in:
Bibliographic Details
Published in2024 IEEE Bangalore Humanitarian Technology Conference (B-HTC) pp. 38 - 42
Main Authors Ananth, Alaka, Bhat, Sachin S, K, Gururaj, D S, Krishna Prasad
Format Conference Proceeding
LanguageEnglish
Published IEEE 22.03.2024
Subjects
Convolution Neural Network
Deep Curve Estimation Network
Enhancement Curves
Estimation
Image color analysis
Image Enhancement
Image quality
Imaging
Loss function
Measurement
Neural networks
Underwater Photography
Visualization
Online AccessGet full text

Cover

More Information
Summary:Underwater imaging poses different challenges to the enhancement techniques because of the distortion, quality, absence of natural light and many other factors. In this paper, we have explored the efficiency of deep learning based algorithms in the realm of image enhancement. Our methodology explores the implementation and optimization of Zero reference Deep Curve Estimation Network to refine the enhancement process. It can be observed that the tailored network significantly improve the visual quality of underwater images by preserving its spatial features. Results are evaluated through both quantitative and qualitative measures which reveal substantial improvement of traditional algorithms. Different loss functions along with the evaluation metrics are used to underscore the potential of the proposed algorithm.
DOI:10.1109/B-HTC60740.2024.10564068