多通道区域建议的多尺度X光安检图像检测

doi:10.3778/j.issn.1002-8331.2008-0345

摘要/Abstract

摘要： 针对安检X光图像检测中的违禁品尺度差异问题，对Faster RCNN网络进行改进，提出一种基于多通道区域建议网络（muiti-channel region proposal network，MCRPN）。考虑到不同层卷积特征在视觉语义上的互补性，进行多层特征提取，融合VGG16高层较丰富的语义特征和低层较浅的边缘特征；修改多通道RPN中的锚框参数，将生成的多尺度候选目标区域分别映射到对应的特征图上，构建多尺度违禁品检测网络；在多通道上引入膨胀卷积，设计一种多分支膨胀卷积模块（dilated convolutions module，DCM），增大感受野，增强不同尺度的特征。将改进的算法在自制数据集SIXray_OD上进行实验，检测的平均精度达到84.69%，测性能较原网络提高了6.28%。实验结果表明，改进算法的识别精度有一定提高。

关键词: 目标检测, Faster RCNN模型, 多尺度, 多通道, 膨胀卷积

Abstract: Aiming at the scale difference problem of prohibited items in X-ray images security inspection, the Faster RCNN network is improved and a multi-channel region proposal network（MCRPN） is designed on the basis of Faster RCNN. Firstly, considering the complementarity of different layers convolutional features in visual semantics, multi-layer feature extraction is introduced to combine richer high-level semantic features and low-level edge features of VGG16. Secondly, the anchor frame parameters in multi-channel RPN are modified. The generated multi-scale candidate regions are mapped to the corresponding feature maps, and a multi-scale contraband detection network is constructed. Finally, dilated convolution is introduced, and a dilated convolution module（DCM） is added to the multi channels to increase the receptive field, and the feature of different scales are enhanced. The improved algorithm is tested on the self-made data set SIXray_OD. The average detection accuracy reaches 84.69%, and the detection performance is improved by 6.28% compared with the original Faster RCNN network. The experimental results show that the recognition accuracy of the proposed algorithm has been improved.

Key words: object detection, Faster RCNN model, multi-scale, multi-channel, dilated convolution

康佳楠, 张良. 多通道区域建议的多尺度X光安检图像检测[J]. 计算机工程与应用, 2022, 58(1): 224-231.

KANG Jianan, ZHANG Liang. Multi-scale X-Ray Security Inspection Image Detection with Multi-channel Region Proposal[J]. Computer Engineering and Applications, 2022, 58(1): 224-231.

参考文献

[1] 卫星，张海涛，陆阳，等.结合并行特征传递深度学习网络的矿井行人检测[J].计算机辅助设计与图形学学报，2019，31（12）：2091-2100.
WEI X，ZHANG H T，LU Y，et al.Mine pedestrian detection with deep learning network of parallel feature transfer[J].Journal of Computer-Aided Design & Computer Graphics，2019，31（12）：2091-2100.
[2] 徐家臻，李婷，杨巍.多尺度局部特征选择的行人重识别算法[J].计算机工程与应用，2020，56（2）：141-145.
XU J Z，LI T，YANG W.Person re-identification by multi-scale local feature selection[J].Computer Engineering and Applications，2020，56（2）：141-145.
[3] 葛宏孔，罗恒利，董佳媛.基于深度学习的非实验场景人脸属性识别[J].计算机科学，2019，46（S2）：246-250.
GE H K，LUO H L，DONG J Y.Face attributes in wild based on deep learning[J].Computer Science，2019，46（S2）：246-250.
[4] GUO G，ZHANG N.A survey on deep learning based face recognition[J].Computer Vision and Image Understanding，2019，189：3077-3142.
[5] XU X，TAO Z，MING W，et al.Intelligent monitoring and diagnostics using a novel integrated model based on deep learning and multi-sensor feature fusion[J].Measurement，2020，165：108086.
[6] PANAYIDES A S，PATTICHIS M，LEANDROU S，et al.Radiogenomics for precision medicine with a big data analytics perspective[J].IEEE Journal of Biomedical & Health Informatics，2019，23（5）：2063-2079.
[7] KAMRAN F，SHAHZAD M，SHAFAIT F.Automated military vehicle detection from low-altitude aerial images[C]//Proceedings of 2018 Digital Image Computing：Techniques and Applications（DICTA），2018：479-483.
[8] 韩欣欣，叶奇玲.基于SIFT和HOG特征融合的人体行为识别方法[J].计算机技术与发展，2019，29（6）：71-74.
HAN X X，YE Q L.Human action recognition based on feature fusion of SIFT and HOG[J].Computer Technology and Development，2019，29（6）：71-74.
[9] GIRSHICK R，DONAHUE J，DARRELL T，et al.Rich feature hierarchies for accurate object detectionand semantic segmentation[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition（CVPR），2014：580-587.
[10] HE K M，ZHANG X Y，REN S Q，et al.Spatial pyramid pooling in deep convolutional network for visual recognition[J].IEEE Transactions on Pattern Analysis and Machine Intelligence 2015，37（9）：1904-1916.
[11] GIRSHICK R.Fast R-CNN[C]//Proceedings of the IEEE International Conference on Computer Vision，2015：176-183.
[12] REN S，HE K，GIRSHICK R，et al.Faster R-CNN：Towards real-time object detection with region proposal networks[C]//Advances in Neural Information Processing Systems，2015：91-99.
[13] LIU W，ANGUELOV D，ERHAN D，et al.SSD：Single shot multibox detector[C]//Proceedings of European Conference on Computer Vision，2016：21-37.
[14] CHA Y L，YANG L R.The three dimensional show of security images from multi-angle X-ray[C]//Proceedings of the 15th National Symposium on Image and Graphics.Beijing：Tsinghua University Press，2010：468-471．
[15] 董浩.基于多视角的X射线图像分析方法研究[D].南京：东南大学，2016：38-49.
DONG H.The research on multiview based X-ray image analysis methods[D].Nanjing：Southeast University，2016：38-49.
[16] RIFFO V，MERY D.Automated detection of threat object using adapted implicit shape model[J].IEEE Transactions on Systems，Man and Cybernetics Systems，2016，46（4）：472-482．
[17] RIFFO V，FLORES S，MERY D.Threat object detection in X-ray imaging using an active vision approach[J].Journal of Nondestructive Evaluation，2017，36（3）：44.
[18] 曲长波，姜思瑶，吴德阳.空洞卷积的多尺度语义分割网络[J].计算机工程与应用，2019，55（24）：91-95.
QU C B，JIANG S Y，WU D Y.Multiscale semantic segmentation network based on cavity convolution[J].Computer Engineering and Applications，2019，55（24）：91-95.
[19] 左静，巴玉林.一种多尺度融合的深度人群计数算法[J/OL].激光与光电子学进展：1-12（2020-06-23）[2020-06-26].http：//kns.cnki.net/kcms/detail/31.1690.TN.20200623.
1440.014.html.
ZUO J，BA Y L.A depth population counting algorithm based on multi-scale Fusion[J/OL].Laser & Optoelectronics Progress：1-12（2020-06-23）[2020-06-26].http：//kns.cnki.net/kcms/detail/31.1690.TN.20200623.1440.014.html.
[20] 孙海铭，时兆峰，李晗，等.基于Faster R-CNN的精密零部件的识别方法[J].飞控与探测，2020，3（2）：26-36.
SUN H M，SHI Z F，LI H，et al.Detection method of precision parts based on faster R-CNN[J].Flight Control and Detection，2020，3（2）：26-36.
[21] 邝先验，李洪伟，杨柳.基于改进Faster R-CNN算法的两轮车视频检测[J].现代电子技术，2020，43（9）：129-134.
KANG X Y，LI H W，YANG L.Detection method of precision parts based on faster R-CNN[J].Modern Electronics Technique，2020，43（9）：129-134.
[22] YU F，KOLTUN V，FUNKHOUSER T.Dilated residual networks[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition（CVPR），2017.