A Retinex based GAN Pipeline to Utilize Paired and Unpaired Datasets for Enhancing Low Light Images

06/27/2020
by   Harshana Weligampola, et al.
0

Low light image enhancement is an important challenge for the development of robust computer vision algorithms. The machine learning approaches to this have been either unsupervised, supervised based on paired dataset or supervised based on unpaired dataset. This paper presents a novel deep learning pipeline that can learn from both paired and unpaired datasets. Convolution Neural Networks (CNNs) that are optimized to minimize standard loss, and Generative Adversarial Networks (GANs) that are optimized to minimize the adversarial loss are used to achieve different steps of the low light image enhancement process. Cycle consistency loss and a patched discriminator are utilized to further improve the performance. The paper also analyses the functionality and the performance of different components, hidden layers, and the entire pipeline.

READ FULL TEXT

page 5

page 6

research
10/03/2021

Semantic-Guided Zero-Shot Learning for Low-Light Image/Video Enhancement

Low-light images challenge both human perceptions and computer vision al...
research
06/17/2019

EnlightenGAN: Deep Light Enhancement without Paired Supervision

Deep learning-based methods have achieved remarkable success in image re...
research
10/11/2021

SurroundNet: Towards Effective Low-Light Image Enhancement

Although Convolution Neural Networks (CNNs) has made substantial progres...
research
07/03/2022

Cycle-Interactive Generative Adversarial Network for Robust Unsupervised Low-Light Enhancement

Getting rid of the fundamental limitations in fitting to the paired trai...
research
09/13/2022

Document Image Binarization in JPEG Compressed Domain using Dual Discriminator Generative Adversarial Networks

Image binarization techniques are being popularly used in enhancement of...
research
12/24/2020

UMLE: Unsupervised Multi-discriminator Network for Low Light Enhancement

Low-light image enhancement, such as recovering color and texture detail...
research
01/13/2022

Learning Enhancement of CNNs via Separation Index Maximizing at the First Convolutional Layer

In this paper, a straightforward enhancement learning algorithm based on...

Please sign up or login with your details

Forgot password? Click here to reset