Low‐light image enhancement via lightweight custom non‐linear transform network

• Published in: Electronics letters Vol. 60; no. 19
• Main Author: Li, Yang
• Format: Journal Article
• Language: English
• Published: Wiley 01.10.2024
• Subjects: image and vision processing and display technology; image enhancement; multimedia computing
• DOI: 10.1049/ell2.70053
• ISSN: 0013-5194

Convolutional neural network (CNN)‐based models have shown significant progress in low light image enhancement. However, many existing models possess a large number of parameters, making them unsuitable for deployment on terminal devices. Moreover, adjustments to brightness, contrast, and colour in images are often non‐linear, and convolution is not the best at capturing complex non‐linear relationships in image data. To address these issues, a model based on an end‐to‐end custom non‐linear transform network (CNTNet) is proposed. CNTNet combines a custom non‐linear transform layer with CNN layers to achieve image contrast and detail enhancement. The CNT layer in this model introduces transformation parameters at multiple scales to manipulate input images within various ranges. CNTNet progressively processes images by stacking multiple non‐linear transform layers and convolutional layers while integrating residual connections to capture and leverage subtle image features. The final output is generated through convolutional layers to obtain enhanced images. Experimental results of CNTNet demonstrate that, while maintaining a comparable level of image quality evaluation metrics to mainstream models, it significantly reduces the parameter count to only 2K. Piecewise linear mapping and convolution are both advantageous tools in image enhancement. This paper designs a network called custom non‐linear transform network, which combines trainable non‐linear mapping with a convolutional neural network. The advantage of trainable nonlinear mapping is its lightweight nature and effective adjustment of image brightness. Therefore, custom non‐linear transform network significantly reduces the number of model parameters while maintaining enhancement performance.