7 March 2024 in IEEE Transactions on Instrumentation and Measurement
Deep-learning-based detection methods have been widely applied to industrial defect inspection. However, directly using vanilla detection methods fails to achieve satisfying performance due to the lack of identifiable features. In this paper, a novel attention-based multi-scale feature fusion method (AMFF) is proposed, aiming to enhance defect features and improve defect identification by leveraging attention mechanism in the feature fusion. AMFF includes self-enhanced attention module (SEAM) and cross-enhanced attention module (CEAM). SEAM is performed on a single feature map, which first adopts multiple dilation convolutions to enrich contextual information without compromising resolution and then utilizes the intra-layer attention on the current feature map. CEAM takes both the current feature map and the adjacent feature map as input to perform cross-layer attention. The adjacent feature map is modulated with the guidance of the current feature map, which is then combined with the current feature map and the output of SEAM for final prediction. AMFF is utilized in current feature fusion networks, e.g., FPN and PAFPN, and is further integrated into prevalent detectors to guide them to pay more attention to defects rather than the background. Extensive experiments are conducted on two real industrial datasets released by Tianchi platform, i.e., fabric and aluminum defect datasets. For each dataset, 500 images are randomly selected for test and the rest for training. The proposed AMFF is demonstrated to significantly boost defect detection accuracy with acceptable computational cost, and the real-time performance could fully satisfy practical requirements.
11 December 2024 in IEEE Access
Deep learning models for object detection have shown favorable results in controlled environments. However, owing to the complexity of underwater environments, these models suffer from target blurring caused by noise, as well as visual disturbances caused by the protective coloration of underwater organisms. These methods are not sufficiently equipped to effectively address the challenge of underwater object detection. To address the challenges in detection caused by underwater blurriness and the camouflage of objects. In this study, a method called Spatial Semantic Encoding Fusion(SSEF) is proposed. Pooling and bilinear interpolation are utilized by SSEF to standardize features across different scales. The standardized features are then fused using the Hadamard product. The fused features are encoded separately for spatial information and channel semantic information through a dual-branch structure. Finally, the dual-branch features are combined by point-wise addition. SSEF is utilized to enhance the multi-scale feature fusion approach of RT-DETR-r18, proposing an underwater object detection framework named Underwater-DETR(UW-DETR). The experiments are conducted using the UTDAC2020 and Brackish datasets. The experimental results show that UW-DETR outperforms other underwater object detection methods and UW-DETR meets the requirements for underwater applications.
1 January 2026 in IEEE Transactions on Instrumentation and Measurement