GAF-Net: A new automated segmentation method based on multiscale feature fusion and feedback module

• Published in: Pattern recognition letters 2025-01, Vol.187, p.86-92
• Main Authors: Wen, Long, Ye, Yuxing, Zuo, Lei
• Format: Article
• Language: English
• Subjects: Convolutional Neural Network; Crack segmentation; Feature fusion and feature feedback; Global feature module
• ISSN: 0167-8655
• DOI: 10.1016/j.patrec.2024.11.025

•A multi-scale feature fusion feedback module is proposed to refine the local features;.•The global feature module is developed to obtain a fine global information map;.•The adaptive feature fusion is proposed to obtain final segmentation map with global feature;.•The multi-level feature maps are deeply supervised to accelerate the network convergence. Surface defect detection (SDD) is the necessary technique to monitor the surface quality of production. However, fine grain defects caused by stress loading, environmental influences, and construction defects is still a challenge to detect. In this research, the convolutional neural network for crack segmentation is developed based on the feature fusion and feedback on the global features and multi-scale feature (GAF-Net). First, a multi-scale feature feedback module (MSFF) is proposed, which uses four different scales to refine local features by fusing high-level and sub-high-level features to perform feedback correction. Secondly, the global feature module (GF) is proposed to generate a fine global information map using local features and adaptive weighted fusion with the correction map for crack detection. Finally, the GAF-Net network with multi-level feature maps is deeply supervised to accelerate GAF-Net and improve the detection accuracy. GAF-Net is trained and experimented on three publicly available pavement crack datasets, and the results show that GAF-Net achieves state-of-the-art results in the IoU segmentation metrics when compared to other deep learning methods (Crackforest: 53.61 %; Crack500: 65.19 %; DeepCrack: 81.63 %).