基于改进Faster R-CNN的小尺度行人检测

doi:10.19678/j.issn.1000-3428.0055817

摘要/Abstract

摘要： 为提高小尺度行人检测的准确性，提出一种基于改进Faster R-CNN的目标检测方法。通过引入基于双线性插值的对齐池化层，避免感兴趣区域池化过程中两次量化操作导致的位置偏差，同时设计基于级联的多层特征融合策略，将具有丰富细节信息的浅层特征图和具有抽象语义信息的深层特征图进行通道叠加，从而解决小尺度行人在深层特征图中特征信息缺乏的问题。在INRIA和PASCAL VOC2012数据集上的实验结果表明，在小尺度行人检测效率相同的情况下，该方法相比基于Faster R-CNN的检测方法平均精确率均值分别提高了17.58%和23.78%。

关键词: 小尺度行人检测, 区域建议网络, 感兴趣区域池化, Faster R-CNN网络, 特征融合

Abstract: To improve the accuracy of small-scale pedestrian detection,this paper proposes a target detection method based on improved Faster R-CNN.The network structure uses a new aligned pooling layer based on bilinear interpolation to avoid the positional deviation caused by two quantization operations in Region of Interest(ROI)pooling.Then a cascade-based multi-layer feature fusion strategy is designed,which concatenates shallow feature maps with rich detail information and deep feature maps with abstract semantic information to address the insufficiency of feature information of small-scale pedestrians in deep feature maps.Experimental results on INRIA and PASCAL VOC2012 datasets show that the proposed method increases the mean Average Precision(mAP) by 17.58% and 23.78% respectively compared with detection method based on Faster R-CNN with the same efficiency of small-scale pedestrian detection.

Key words: small-scale pedestrian detection, Region Proposal Network(RPN), Region of Interest(ROI) pooling, Faster R-CNN network, feature fusion

中图分类号:

TP391

陈泽, 叶学义, 钱丁炜, 魏阳洋. 基于改进Faster R-CNN的小尺度行人检测[J]. 计算机工程, 2020, 46(9): 226-232,241.

CHEN Ze, YE Xueyi, QIAN Dingwei, WEI Yangyang. Small-Scale Pedestrian Detection Based on Improved Faster R-CNN[J]. Computer Engineering, 2020, 46(9): 226-232,241.

http://www.ecice06.com/CN/Y2020/V46/I9/226

图/表 12

20200915141600

20200915141603

20200915141605

20200915141608

20200915141611

20200915141614

20200915141617

20200915141620

20200915141624

20200915141627

20200915141630

20200915141634

参考文献

[1] ZHANG S S,BENENSON R,SCHIELE B,et al.CityPersons:a diverse dataset for pedestrian detection[C]//Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C.,USA:IEEE Press,2017:4457-4465.
[2] XUE Lixia,ZHONG Xin,WANG Ronggui,et al.Mid-low resolution vehicle type recognition based on deep feature fusion[J].Computer Engineering,2019,45(1):233-238,245.(in Chinese) 薛丽霞,钟欣,汪荣贵,等.基于深度特征融合的中低分辨率车型识别[J].计算机工程,2019,45(1):233-238,245.
[3] LI Hongyan,LI Chungeng,AN Jubai,et al.Attention mechanism improves CNN remote sensing[J].Journal of Image and Graphics,2019,24(8):1400-1408.(in Chinese) 李红艳,李春庚,安居白,等.注意力机制改进卷积神经网络的遥感图像目标检测[J].中国图象图形学报,2019,24(8):1400-1408.
[4] REN S,GIRSHICK R,GIRSHICK R,et al.Faster R-CNN:towards real-time object detection with region proposal networks[J].IEEE Transactions on Pattern Analysis & Machine Intelligence,2017,39(6):1137-1149.
[5] HE Kaiming,ZHANG Xiangyu,REN Shaoqing,et al.Spatial pyramid pooling in deep convolutional networks for visual recognition[J].IEEE Transactions on Pattern Analysis & Machine Intelligence,2014,37(9):1904-1916.
[6] GIRSHICK R.Fast R-CNN[EB/OL].[2019-05-10].https://arxiv.org/pdf/1504.08083.pdf.
[7] DAI Jifeng,LI Yi,HE Kaiming,et al.R-FCN:object detection via region-based fully convolutional networks[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C.,USA:IEEE Press,2016:379-387.
[8] REDMON J,FARHADI A.YOLOv3:an incremental improvement[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C.,USA:IEEE Press,2018:1-6.
[9] JIAO Tianchi,LI Qiang,LIN Maosong,et al.Target detection method combining inverted residual block and YOLOv3[J].Transducer and Microsystem Technologies,2019,38(9):144-146,156.(in Chinese) 焦天驰,李强,林茂松,等.结合反残差块和YOLOv3的目标检测法[J].传感器与微系统,2019,38(9):144-146,156.
[10] LIU W,ANGUELOV D,ERHAN D,et al.SSD:single shot multibox detector[C]//Proceedings of the 14th European Conference on Computer Vision.Amsterdam,the Netherlands:[s.n.],2016:21-37.
[11] LIN T Y,GOYAL P,GIRSHICK R,et al.Focal loss for dense object detection[C]//Proceedings of 2017 IEEE International Conference on Computer Vision.Washington D.C.,USA:IEEE Press,2017:2999-3007.
[12] ZHANG Liliang,LIN Liang,LIANG Xiaodan,et al.Is Faster R-CNN doing well for pedestrian detection?[C]//Proceedings of European Conference on Computer Vision.Berlin,Germany:Springer,2016:443-457.
[13] LI Jianan,LIANG Xiaodan,SHEN Shengmei,et al.Scale-aware Fast R-CNN for pedestrian detection[C]//Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C.,USA:IEEE Press,2017:1-6.
[14] CAI Z W,FAN Q F,FERIS R S,et al.A unified multi-scale deep convolutional neural network for fast object detection[C]//Proceedings of the 14th European Conference on Computer Vision.Amsterdam,the Netherlands:[s.n.],2016:354-370.
[15] MAO Jiayuan,XIAO Tete,JIANG Yuning,et al.What can help pedestrian detection?[C]//Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C.,USA:IEEE Press,2017:1-5.
[16] LI Chengyang,SONG Dan,TONG Ruofeng,et al.Illumination-aware Faster R-CNN for robust multispectral pedestrian detection[EB/OL].[2019-05-10].https://arxiv.org/pdf/1803.05347.pdf.
[17] LIU Songtao,HUANG Di,WANG Yunhong.Adaptive NMS:refining pedestrian detection in a crowd[EB/OL].[2019-05-10].https://arxiv.org/pdf/1904.03629v1.pdf.
[18] KULKARNI A,CALLAN J.Selective search[J].ACM Transactions on Information Systems,2015,33(4):1-33.
[19] GIRSHICK R,DONAHUE J,DARRELLAND T,et al.Rich feature hierarchies for object detection and semantic segmentation[C]//Proceedings of 2014 IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C.,USA:IEEE Press,2014:1-5.
[20] SIMONYAN K,ZISSERMAN A.Very deep convolutional networks for large-scale image recognition[EB/OL].[2019-05-10].https://arxiv.org/pdf/1409.1556.pdf.
[21] HE K,GKIOXARI G,DOLLAR P,et al.Mask R-CNN[C]//Proceedings of 2017 IEEE International Conference on Computer Vision.Washington D.C.,USA:IEEE Press,2017:1-12.
[22] SHI W,CABALLERO J,THEIS L,et al.Is the deconvolution layer the same as a convolutional layer?[EB/OL].[2019-05-10].https://arxiv.org/ftp/arxiv/papers/1609/1609.07009.pdf.
[23] IOFFE S,SZEGEDY C.Batch normalization:accelerating deep network training by reducing internal covariate shift[C]//Proceedings of International Conference on International Conference on Machine Learning.Miami,USA:[s.n.],2015:1-11.
[24] BODLA N,SINGH B,CHELLAPPA R,et al.Soft-NMS:improving object detection with one line of code[C]//Proceedings of 2017 IEEE International Conference on Computer Vision.Washington D.C.,USA:IEEE Press,2017:5562-5570.

选择文件类型/文献管理软件名称

选择包含的内容