Low-Light Image Enhancement with Normalizing Flow

Yufei Wang, Renjie Wan*, Wenhan Yang, Haoliang Li, Lap-Pui Chau, Alex C. Kot

*Corresponding author for this work

Research output: Chapter in book/report/conference proceedingConference contributionpeer-review

Abstract

To enhance low-light images to normally-exposed ones is highly ill-posed, namely that the mapping relationship between them is one-to-many. Previous works based on the pixel-wise reconstruction losses and deterministic processes fail to capture the complex conditional distribution of normally exposed images, which results in improper brightness, residual noise, and artifacts. In this paper, we investigate to model this one-to-many relationship via a proposed normalizing flow model. An invertible network that takes the low-light images/features as the condition and learns to map the distribution of normally exposed images into a Gaussian distribution. In this way, the conditional distribution of the normally exposed images can be well modeled, and the enhancement process, i.e., the other inference direction of the invertible network, is equivalent to being constrained by a loss function that better describes the manifold structure of natural images during the training. The experimental results on the existing benchmark datasets show our method achieves better quantitative and qualitative results, obtaining better-exposed illumination, less noise and artifact, and richer colors.
Original languageEnglish
Title of host publication36th AAAI Conference on Artificial Intelligence, Acceptd Papers
PublisherAAAI press
Number of pages9
Publication statusPublished - Feb 2022
Event36th AAAI Conference on Artificial Intelligence, AAAI 2022 - Virtual
Duration: 22 Feb 20221 Mar 2022
https://aaai.org/Conferences/AAAI-22/

Conference

Conference36th AAAI Conference on Artificial Intelligence, AAAI 2022
Period22/02/221/03/22
Internet address

Fingerprint

Dive into the research topics of 'Low-Light Image Enhancement with Normalizing Flow'. Together they form a unique fingerprint.

Cite this