Adaptive Exposure Correction in Low-Light Images Using Frequency Analysis

DOI

10.2991/978-94-6463-870-7_10How to use a DOI?

Keywords

Low illumination image; Illumination- reflection model; Deep Learning-based Restoration; Extreme Low-Light Conditions; Adaptive Exposure Correction; Parallel Multi-Scale Processing; Real-Time Image Processing; Residual Dense Block (RDB)

Abstract

The effort of recovering images in extremely low-light environments is due to the occurrence of high levels of noise, color distortions, and low contrast. Even though deep learning-based techniques have been highly successful, the majority of current tactics prioritize restoration quality over computational complexity and memory usage, making them unsuitable for real-time applications. We suggest a fast and lightweight deep learning framework that efficiently restores extreme low-light images while maintaining high visual quality to address this limitation. By utilizing parallel multi- scale processing, the proposed model ensures that most operations are performed at higher scale-spaces, which reduces computational overhead and inference time. Also, we include an adaptive amplifier module that computes the most suitable amplification factor directly from input image, eliminating need for ground truth exposure values and making this technique valid in real-life scenarios. The design of our architecture is optimized for real-time performance, enabling the restoration of 4K resolution images in 1 s on a CPU and achieving 32-bit refresh times. FPS on GPU. Our model is quicker, inexpensive in terms of computational power, and requires fewer model parameters while maintaining comparable restoration quality compared to current systems. And our model does great at dealing with a wide range of cameras and does great object detection and other tasks without any extra refinements. Detailed analyses indicate that our approach is greater to existing lightweight systems in terms of efficiency and perceptual quality. The use of our framework for real- time, scaled up (by adding an adaptive amplification mechanism) and high-quality scaling is the foundation for practical refurbishment in extreme low-light images that are ideal for applications like observation systems or night photography.

Copyright

© 2025 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - R. K. Supriya
AU  - S. R. Sowmya
AU  - Podamala Rahul Sai
AU  - Khushi Manjunath
AU  - S. K. Pooja
AU  - Preetham Srinivasa
PY  - 2025
DA  - 2025/10/22
TI  - Adaptive Exposure Correction in Low-Light Images Using Frequency Analysis
BT  - Proceedings of the International Conference on Smart Innovations in Electrical Engineering (ICSIEE 2025)
PB  - Atlantis Press
SP  - 86
EP  - 97
SN  - 2352-5401
UR  - https://doi.org/10.2991/978-94-6463-870-7_10
DO  - 10.2991/978-94-6463-870-7_10
ID  - Supriya2025
ER  -