Material-Aware Path Aggregation Network and Shape Decoupled SIoU for X-ray Contraband Detection

Xiang, Nan; Gong, Zehao; Xu, Yi; Xiong, Lili

doi:10.3390/electronics12051179

Cited by 7 publications

(3 citation statements)

References 56 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Li et al improved the YOLOv5 model by compressing channels, optimizing parameters, and proposing a new YOLO-FIRI model for infrared target detection problems such as low recognition rates and high false alarm rates due to long distances, weak energy, and low resolutions [35]. Xiang et al integrated both MCA and SCA modules into the YOLOx framework, enabling the acquisition of material information for contraband while expanding the model's receptive field, thereby enhancing the detection efficiency [36]. These improvements have significantly impacted the detection quality of contraband detection algorithms.…”

Section: Deep Learning For Contraband X-ray Image Detectionmentioning

confidence: 99%

YOLO-CID: Improved YOLOv7 for X-ray Contraband Image Detection

Gan,

Wan,

Lei

et al. 2023

Electronics

View full text Add to dashboard Cite

Currently, X-ray inspection systems may produce false detections due to factors such as the varying sizes of contraband images, complex backgrounds, and blurred edges. To address this issue, we propose the YOLO-CID method for contraband image detection. Firstly, we designed the MP-OD module in the backbone network to enhance the model’s ability to extract key information from complex background images. Secondly, at the neck of the network, we designed a simplified version of BiFPN to add cross-scale connection lines in the feature fusion structure, to preserve deeper semantic information and enhance the network’s ability to represent objects in low-contrast or occlusion situations. Finally, we added a new object detection layer to improve the model’s accuracy in detecting small objects in dense environments. Experimental results on the PIDray public dataset show that the average accuracy rate of the YOLO-CID algorithm is 82.7% and the recall rate is 81.2%, which are 4.9% and 3.2% higher than the YOLOv7 algorithm, respectively. At the same time, the mAP on the CLCXray dataset reached 80.2%. Additionally, it can achieve a real-time detection speed of 40 frames per second and 43 frames per second in real scenes. These results demonstrate the effectiveness of the YOLO-CID algorithm in X-ray contraband detection.

show abstract

Section: Deep Learning For Contraband X-ray Image Detectionmentioning

confidence: 99%

YOLO-CID: Improved YOLOv7 for X-ray Contraband Image Detection

Gan,

Wan,

Lei

et al. 2023

Electronics

View full text Add to dashboard Cite

show abstract

“…The mean accuracies of the improved algorithm reached 88.2%, 89.2%, and 70.5%, respectively, and improved by 2.5%, 1.2%, and 5.0%, respectively, on the three test sets of the PIDray datasets, respectively. Xiang et al [22] designed a new network integrated MSA (multi-scale smoothed atrous) convolution and MCA (material-aware coordinate attention) module based on YOLOX algorithm, and introduced the SD-SIoU (shape decoupled SIoU) loss function for learning the prohibited items shape features. The improved model reached the accuracies of 92.92% and 91.1% on the OPIXray [23] datasets and SIXray datasets, respectively.…”

Section: Introductionmentioning

confidence: 99%

Lightweight Detection Method for X-ray Security Inspection with Occlusion

Wang,

Shi

et al. 2024

Sensors

View full text Add to dashboard Cite

Identifying the classes and locations of prohibited items is the target of security inspection. However, X-ray security inspection images with insufficient feature extraction, imbalance between easy and hard samples, and occlusion lead to poor detection accuracy. To address the above problems, an object-detection method based on YOLOv8 is proposed. Firstly, an ASFF (adaptive spatial feature fusion) and a weighted feature concatenation algorithm are introduced to fully extract the scale features from input images. In this way, the model can learn further details in training. Secondly, CoordAtt (coordinate attention module), which belongs to the hybrid attention mechanism, is embedded to enhance the learning of features of interest. Then, the slide loss function is introduced to balance the simple samples and the difficult samples. Finally, Soft-NMS (non-maximum suppression) is introduced to resist the conditions containing occlusion. The experimental result shows that mAP (mean average precision) achieves 90.2%, 90.5%, 79.1%, and 91.4% on the Easy, Hard, and Hidden sets of the PIDray and SIXray public test set, respectively. Contrasted with original model, the mAP of our proposed YOLOv8n model increased by 2.7%, 3.1%, 9.3%, and 2.4%, respectively. Furthermore, the parameter count of the modified YOLOv8n model is roughly only 3 million.

show abstract

“…Finally, they added a detection head with inputs of low-level and high-resolution feature maps to improve the model’s ability to detect small objects. Xiang et al [ 12 ] proposed an enhanced network architecture to address the issue of severe object occlusion in X-ray prohibited item detection. This enhanced structure comprised multiscale smoothed atrous convolutions and a material-aware coordinate attention module.…”

Section: Introductionmentioning

confidence: 99%

ScanGuard-YOLO: Enhancing X-ray Prohibited Item Detection with Significant Performance Gains

Huang,

Zhang

2023

Sensors

View full text Add to dashboard Cite

To address the problem of low recall rate in the detection of prohibited items in X-ray images due to the severe object occlusion and complex background, an X-ray prohibited item detection network, ScanGuard-YOLO, based on the YOLOv5 architecture, is proposed to effectively improve the model’s recall rate and the comprehensive metric F1 score. Firstly, the RFB-s module was added to the end part of the backbone, and dilated convolution was used to increase the receptive field of the backbone network to better capture global features. In the neck section, the efficient RepGFPN module was employed to fuse multiscale information from the backbone output. This aimed to capture details and contextual information at various scales, thereby enhancing the model’s understanding and representation capability of the object. Secondly, a novel detection head was introduced to unify scale-awareness, spatial-awareness, and task-awareness altogether, which significantly improved the representation ability of the object detection heads. Finally, the bounding box regression loss function was defined as the WIOUv3 loss, effectively balancing the contribution of low-quality and high-quality samples to the loss. ScanGuard-YOLO was tested on OPIXray and HiXray datasets, showing significant improvements compared to the baseline model. The mean average precision (mAP@0.5) increased by 2.3% and 1.6%, the recall rate improved by 4.5% and 2%, and the F1 score increased by 2.3% and 1%, respectively. The experimental results demonstrate that ScanGuard-YOLO effectively enhances the detection capability of prohibited items in complex backgrounds and exhibits broad prospects for application.

show abstract

Material-Aware Path Aggregation Network and Shape Decoupled SIoU for X-ray Contraband Detection

Cited by 7 publications

References 56 publications

YOLO-CID: Improved YOLOv7 for X-ray Contraband Image Detection

YOLO-CID: Improved YOLOv7 for X-ray Contraband Image Detection

Lightweight Detection Method for X-ray Security Inspection with Occlusion

ScanGuard-YOLO: Enhancing X-ray Prohibited Item Detection with Significant Performance Gains

Contact Info

Product

Resources

About