Electrical Engineering and Systems Science > Image and Video Processing
[Submitted on 22 Apr 2021 (v1), last revised 21 May 2021 (this version, v2)]
Title:Frequency Domain Loss Function for Deep Exposure Correction of Dark Images
View PDFAbstract:We address the problem of exposure correction of dark, blurry and noisy images captured in low-light conditions in the wild. Classical image-denoising filters work well in the frequency space but are constrained by several factors such as the correct choice of thresholds, frequency estimates etc. On the other hand, traditional deep networks are trained end-to-end in the RGB space by formulating this task as an image-translation problem. However, that is done without any explicit constraints on the inherent noise of the dark images and thus produce noisy and blurry outputs. To this end we propose a DCT/FFT based multi-scale loss function, which when combined with traditional losses, trains a network to translate the important features for visually pleasing output. Our loss function is end-to-end differentiable, scale-agnostic, and generic; i.e., it can be applied to both RAW and JPEG images in most existing frameworks without additional overhead. Using this loss function, we report significant improvements over the state-of-the-art using quantitative metrics and subjective tests.
Submission history
From: Ojasvi Yadav [view email][v1] Thu, 22 Apr 2021 04:13:40 UTC (4,450 KB)
[v2] Fri, 21 May 2021 12:41:57 UTC (4,506 KB)
Current browse context:
eess.IV
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.