Occluded prohibited object detection in X-ray images with global Context-aware Multi-Scale feature Aggregation

Ma, Chunjie; Li, Zhuo; Li, Jiafeng; Zhang, Yutong; Zhang, Jing

doi:10.1016/j.neucom.2022.11.034

Cited by 23 publications

(11 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…PAA [34] and YOLOv6 [29]. The prohibited object detection methods in X-ray images include MFFNet [40]，FAB [39]，MCIA [13]，DOAM [57]，LA [42]，FA [58]，EAOD-Net [12], Ma et al [59] and POD [38]。 For SIXray dataset, the IoU threshold is set as 0.5 to calculate the mAP. For CLCXray, COCO evaluation criteria is adopted.…”

Section: Comparison With the State-of-the-arts Methodsmentioning

confidence: 99%

See 1 more Smart Citation

CTA-FPN: Channel-Target Attention Feature Pyramid Network for Prohibited Object Detection in X-ray Images

Zhang

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

Fast and accurate prohibited object detection in X-ray images is great challenging. Based on YOLOv6 object detection framework, in this paper, Channel-Target Attention Feature Pyramid Network (CTA-FPN) is proposed for prohibited object detection in X-ray images. It includes two key components: TAAM (Target Aware Attention Module) and CAM (Channel Attention Module). TAAM is to generate the target attention map to enhance the features of prohibited object regions and suppress those of the background regions, so as to solve the problems of object occlusion and cluttered background in X-ray images. CAM is to highlight the feature channels important to the detection tasks, and suppress the irrelevant ones. The target-wise and channel-wise feature enhancement can effectively strengthen the feature representation capability of the network. The proposed CTA-FPN is incorporated into S, M and L models of YOLOv6 respectively, obtaining three X-ray prohibited object detection models. The experimental results on two publicly available benchmark datasets of SIXray and CLCXray show that, CTA-FPN can effectively improve the detection performance of YOLOv6. Especially, YOLOv6-CTA-FPN-L can achieve the state-of-the-arts detection accuracy.

show abstract

Section: Comparison With the State-of-the-arts Methodsmentioning

confidence: 99%

“…Ma el at. [38] proposed a feature enhancement strategy for prohibited object detection. They introduced Gabor convolution and designed a SA (Spatial Attention) mechanism to enhance the network's capability to extract object edge contour information.…”

Section: X-ray Prohibited Object Detectionmentioning

confidence: 99%

CTA-FPN: Channel-Target Attention Feature Pyramid Network for Prohibited Object Detection in X-ray Images

Zhang

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…After that, the feature map f is subjected to a 1 1  convolution operation following the initial width and height so that the feature maps h f and w f with the same dimension as the initial ones are obtained. Then, they are processed by the Sigmoid activation function to obtain the attention weights h g and w g on the width and height directions, respectively, as shown in Equation (3).…”

Section: Coordinate Attentionmentioning

confidence: 99%

Design of security image contraband detection system based on PP-YOLOE+_DCS

Chen,

Jin,

2023

Seventh International Conference on Mechatronics and Intelligent Robotics (ICMIR 2023)

View full text Add to dashboard Cite

Traditional security screening methods mainly use manual identification of security images, there is a low identification of inefficiency, high error of judgement rate,which has become a bottleneck limiting public safety and security. Therefore, to deal with this problem, this paper proposes the security image contraband detection model PP-YOLOE+_DCS, which makes three main improvements on the basis of the PP-YOLOE+ model: To begin with, we introduced deformable convolution within the backbone network that strengthen the models' feature extraction capability; Secondly, we introduced a coordinated attention mechanism among the backbone network and the detection neck for better focusing the model on the object region;Finally, we replaced the original GIOU loss function with the SIOU loss function to improve the detection accuracy and training speed. The improved PP_YOLOE+_DCS model obtained achieves 91.4% detection accuracy,2.8% improvement compared with the baseline model mAP, only 0.24M additional parameters,and 420.2ms inference delay on embedded devices, which provides a new solution for the intelligence of contraband detection.

show abstract

“…XMC R-CNN consists of a material classification algorithm and an organic-inorganic separation algorithm such that two modules can mitigate the accuracy degradation, which is caused by the occlusion problem effectively. Ma et al 9 performed different processing methods for low-frequency and high-frequency information. It can extract multiscale global contextual information by global context feature extraction module from high-frequency and low-frequency information, such as edges and contours, by learnable Gabor convolution layer.…”

Section: Introductionmentioning

confidence: 99%

“…XMC R-CNN consists of a material classification algorithm and an organic–inorganic separation algorithm such that two modules can mitigate the accuracy degradation, which is caused by the occlusion problem effectively. Ma et al 9 . performed different processing methods for low-frequency and high-frequency information.…”

Section: Introductionmentioning

confidence: 99%

Material-aware multiscale atrous convolutional network for prohibited items detection in x-ray image

et al. 2023

View full text Add to dashboard Cite

To resolve the problem of occlusion of the depth information of x-ray images and the detection of small-scale contraband in the detection of contraband objects, an improved prohibited item detection network has been proposed based on YOLOX. First, a material-aware atrous convolution module (MACM) is added to the feature pyramid network to enhance the model's multiscale fusion and extraction ability for material information in x-ray image. Second, a spatial pyramid split attention mechanism (SPSA) is proposed to fuse spatial and channel attention for different scale spatial information features. Finally, CutMix data augmentation strategy is adopted to improve the robustness of the model. The overall performance identification experiments were conducted on the publicly available OPIXray dataset. The average accuracy (mean average precision, mAP) of the method is 93.10%. Compared with the baseline model YOLOX, the mAP is improved by 3.25%. The experimental results show that our method achieves stateof-the-art detection accuracy compared with existing methods.

show abstract

Occluded prohibited object detection in X-ray images with global Context-aware Multi-Scale feature Aggregation

Cited by 23 publications

References 32 publications

CTA-FPN: Channel-Target Attention Feature Pyramid Network for Prohibited Object Detection in X-ray Images

CTA-FPN: Channel-Target Attention Feature Pyramid Network for Prohibited Object Detection in X-ray Images

Design of security image contraband detection system based on PP-YOLOE+_DCS

Material-aware multiscale atrous convolutional network for prohibited items detection in x-ray image

Contact Info

Product

Resources

About