.png){kind=icon}
Return
Enhancing low-light images with optimal graph convolutional neural network using osprey optimization algorithm
.png){kind=icon}
PRE
AI
DOI：10.1007/s12597-026-01198-8.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
A computer vision task known as Low-Light Image Enhancement focuses on restoring and improving image quality captured under poor illumination conditions. The objective is to amplify visibility while suppressing noise, artifacts, and structural distortions, thereby producing images with enhanced luminance, improved clarity, and better perceptual fidelity. This study introduces a novel enhancement strategy that preserves critical scene characteristics and adaptively regulates illumination using an Optimized Graph Convolutional Neural Network (OGCNN) driven by an Osprey Optimization Algorithm (OOA). Initially, the input image is given to the preprocessing stage, where it is rotated by an angle. An Optimal Graph Convolutional Neural Network (OGCNN) is employed to identify image enhancement in low light. Osprey Optimization Algorithm (OOA) optimizes the hyperparameters in Optimal Graph Convolutional Neural Network (OGCNN). Utilizing the MATLAB computing language, the proposed model is implemented. Techniques for improving low-light images are evaluated using datasets LIME (low-light image enhancement), DICM (digital cameras), and MEF (multi-exposure image fusion). In all three datasets, the results clearly show that the suggested model has the minimum Lightness Order Error (LOE) and maximum Structural Similarity Index (SSIM) when compared to the current technique.
Keywords:
Low light image enhancement
Graph convolutional neural network
Osprey optimization algorithm
Hyperparameters
Rotation
Lightness order error
Structural similarity index
Journal
O
IF:
1.8
Papers:
182
Citations:
1.0K
Organization
No organization information available
Cited Papers
No cited papers available