Computer Science ›› 2022, Vol. 49 ›› Issue (8): 136-142.doi: 10.11896/jsjkx.220100132

• Computer Graphics & Multimedia • Previous Articles     Next Articles

Incremental Object Detection Method Based on Border Distance Measurement

LIU Dong-mei, XU Yang, WU Ze-bin, LIU Qian, SONG Bin, WEI Zhi-hui   

  1. School of Computer Science and Engineering,Nanjing University of Science and Technology,Nanjing 210094,China
  • Received:2022-01-17 Revised:2022-02-17 Published:2022-08-02
  • About author:LIU Dong-mei,born in 1996,postgra-duate.Her main research interests include image processing and deep lear-ning.
    XU Yang,born in 1990,Ph.D,associate professor,is a member of China Computer Federation.His main research interests include image processing and machine learning.
  • Supported by:
    National Natural Science Foundation of China(61772274,62071233,61971223,61976117),Natural Science Foundation of Jiangsu Province(BK20211570,BK20180018,BK20191409),Fundamental Research Funds for the Central Universities(30917015104,30919011103,30919011402,30921011209) and China Postdoctoral Science Foundation(2017M611814,2018T110502).

Abstract: Incremental learning has achieved good results in image classification,but it is challenging to apply incremental learning to multi-class object detection.Object detection is more complex than image classification,which combines classification and border regression.At present,the most advanced incremental object detectors adopt the external fixed region suggestion method based on knowledge distillation,which consumes a lot of time and cost.For single-stage detectors,due to the lack of annotation and region advice information for the old class,old objects are usually identified by the detector as the background,resulting in catastrophic forgetting.In this paper,a label selection algorithm based on border distance metric is proposed.It uses the detection results of the old model and the existing dataset labels to select and merge by measuring the coincidence of the bounding boxes,making up for the lack of annotations of the old objects in the new dataset and alleviating catastrophic forgetting.In addition,a module that combines the attention module with the residual module is designed to extract discriminative features at different depths in feature extraction network,to further improve the detection accuracy of model.The proposed method is implemented in the single-stage detection framework,and the effectiveness of the method is verified on PASCAL VOC dataset.Compared with the best model at present,the average accuracy value of the old object and all objects improves by 2.8% and 2.1%,respectively.The pseudo-labels obtained by the proposed method greatly alleviate the forgetting problem,and the attention residual module improves the detection accuracy of the model.

Key words: Attention module, Catastrophic forgetting, Incremental learning, Label selection, Object detection, Pseudo label

CLC Number: 

  • TP391
[1]HE K,ZHANG X,REN S,et al.Spatial Pyramid Pooling inDeep Convolutional Networks for Visual Recognition[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2015,37(9):1904-1916.
[2]REN S,HE K,GIRRSHICK R,et al.Faster R-CNN:Towards Real-Time Object Detection with Region Proposal Networks[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2017,39(6):1137-1149.
[3]LIN T Y,DOLLAR P,GIRSHICK R,et al.Feature Pyramid Networks for Object Detection[C]//IEEE Conference on Computer Vision and Pattern Recognition(CVPR).IEEE Computer Society,2017:2117-2125.
[4]REDMON J,DIVVALA S,GIRSHICK R,et al.You Only Look Once:Unified,Real-time Object Detection[C]//Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition.IEEE Computer Society,2016:779-788.
[5]REDMON J,FARHADI A.YOLOv3:An Incremental Improvement[J].arXiv:1804.02767,2018.
[6]ZHOU X,WANG D,KRÄHENBÜHL P.Objects as Points[J].arXiv:1904.07850,2019.
[7]TIAN Z,SHEN C,CHEN H,et al.FCOS:Fully Convolutional One-Stage Object Detection[C]//IEEE International Conference on Computer Vision.2019:9626-9635.
[8]SHMELKOV K,SCHMID C,ALAHARI K.Incremental Lear-ning of Object Detectors without Catastrophic Forgetting[C]//Proceedings of the IEEE International Conference on Computer Vision.2017:3420-3429.
[9]LI Z,HOIEM D.Learning without Forgetting[J].IEEE Transa-ctions on Pattern Analysis and Machine Intelligence,2017,40(12):2935-2947.
[10]LI D,TASCI S,GHOSH S,et al.RILOD:Near Real-time Incremental Learning for Object Detection at the Edge[C]//Procee-dings of the 4th ACM.New York,2019:113-126.
[11]ZHANG J,ZHANG J,GHOSH S,et al.Class-incrementalLearning via Deep Model Consolidation[C]//IEEE Winter Conference on Applications of Computer Vision.2020:1131-1140.
[12]ZHENG Z,WANG P,REN D,et al.Enhancing Geometric Factors in Model Learning and Inference for Object Detection and Instance Segmentation[EB/OL].
[13]DHAR P,SINGH R V,PENG K C,et al.Learning withoutMemorizing[C]//IEEE Conference on Computer Vision and Pattern Recognition.2019:5138-5146.
[14]WOO S,PARK J,LEE J Y,et al.CBAM:Convolutional Block Attention Module[C]//Proceedings of the European Conference on Computer Vision.2018:3-19.
[15]LIU L,KUANG Z,CHEN Y,et al.IncDet:In Defense of Elastic Weight Consolidation for Incremental Object Detection[J].IEEE Transactions on Neural Networks and Learning Systems,2021,32(6):2306-2319.
[1] WEI Kai-xuan, FU Ying. Re-parameterized Multi-scale Fusion Network for Efficient Extreme Low-light Raw Denoising [J]. Computer Science, 2022, 49(8): 120-126.
[2] WANG Can, LIU Yong-jian, XIE Qing, MA Yan-chun. Anchor Free Object Detection Algorithm Based on Soft Label and Sample Weight Optimization [J]. Computer Science, 2022, 49(8): 157-164.
[3] CHEN Yong-ping, ZHU Jian-qing, XIE Yi, WU Han-xiao, ZENG Huan-qiang. Real-time Helmet Detection Algorithm Based on Circumcircle Radius Difference Loss [J]. Computer Science, 2022, 49(6A): 424-428.
[4] WU Lin, SUN Jing-yu. Multi-branch RA Capsule Network and Its Application in Image Classification [J]. Computer Science, 2022, 49(6): 224-230.
[5] CHEN Jia-zhou, ZHAO Yi-bo, XU Yang-hui, MA Ji, JIN Ling-feng, QIN Xu-jia. Small Object Detection in 3D Urban Scenes [J]. Computer Science, 2022, 49(6): 238-244.
[6] HU Fu-yuan, WAN Xin-jun, SHEN Ming-fei, XU Jiang-lang, YAO Rui, TAO Zhong-ben. Survey Progress on Image Instance Segmentation Methods of Deep Convolutional Neural Network [J]. Computer Science, 2022, 49(5): 10-24.
[7] XU Tao, CHEN Yi-ren, LYU Zong-lei. Study on Reflective Vest Detection for Apron Workers Based on Improved YOLOv3 Algorithm [J]. Computer Science, 2022, 49(4): 239-246.
[8] SHEN Shao-peng, MA Hong-jiang, ZHANG Zhi-heng, ZHOU Xiang-bing, ZHU Chun-man, WEN Zuo-cheng. Three-way Drift Detection for State Transition Pattern on Multivariate Time Series [J]. Computer Science, 2022, 49(4): 144-151.
[9] XU Hua-jie, QIN Yuan-zhuo, YANG Yang. Scene Recognition Method Based on Multi-level Feature Fusion and Attention Module [J]. Computer Science, 2022, 49(4): 209-214.
[10] ZHAO Yue, YU Zhi-bin, LI Yong-chun. Cross-attention Guided Siamese Network Object Tracking Algorithm [J]. Computer Science, 2022, 49(3): 163-169.
[11] XIE Yu, YANG Rui-ling, LIU Gong-xu, LI De-yu, WANG Wen-jian. Human Skeleton Action Recognition Algorithm Based on Dynamic Topological Graph [J]. Computer Science, 2022, 49(2): 62-68.
[12] YUAN Lei, LIU Zi-yan, ZHU Ming-cheng, MA Shan-shan, CHEN Lin-zhou-ting. Improved YOLOv3 Remote Sensing Target Detection Based on Improved Dense Connection and Distributional Ranking Loss [J]. Computer Science, 2021, 48(9): 168-173.
[13] ZHOU Wen-hui, SHI Min, ZHU Deng-ming, ZHOU Jun. Seismic Data Super-resolution Method Based on Residual Attention Network [J]. Computer Science, 2021, 48(8): 24-31.
[14] GONG Hao-tian, ZHANG Meng. Lightweight Anchor-free Object Detection Algorithm Based on Keypoint Detection [J]. Computer Science, 2021, 48(8): 106-110.
[15] QING Lai-yun, ZHANG Jian-gong, MIAO Jun. Temporal Modeling for Online Anomaly Detection [J]. Computer Science, 2021, 48(7): 206-212.
Full text



No Suggested Reading articles found!