Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

GPNet: Key Point Generation Auxiliary Network for Object Detection

GPNet: Key Point Generation Auxiliary Network for Object Detection For existing object detectors, anchor‐based detectors lack global information, while anchor‐free detectors based on key points lack prior position information. The above issues may lead to the imbalance of detection accuracy and shape robustness. In order to alleviate the above contradictions, a new auxiliary network key point generation network (GPNet) is proposed to improve the performance of existing object detectors. Specifically, a series of key points are generated through ground truth (GT) supervision to obtain more global information. These key points generate pseudo boxes (Pbox) with learnable parameters. Pbox has more specific prior information than the manually designed candidate boxes. The scale information of the Pbox is embed into the classification branch to obtain a more appropriate receptive field. In addition, a novel improved strategy for label assignment by combining Pbox and GT to enhance the ability to classify positive and negative samples is proposed. Extensive experiments on multiple dense prediction methods validate the effectiveness of GPNet, with a performance improvement of 1.5 AP over baseline. In particular, with ResNext‐101‐64× 4d‐DCN as the backbone, this method achieves 49.5 AP with single‐scale testing. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Advanced Theory and Simulations Wiley

GPNet: Key Point Generation Auxiliary Network for Object Detection

Download PDF
10 pages

Loading next page...
 
/lp/wiley/gpnet-key-point-generation-auxiliary-network-for-object-detection-ZeWKEcd2jZ
Publisher
Wiley
Copyright
© 2023 Wiley‐VCH GmbH
eISSN
2513-0390
DOI
10.1002/adts.202200894
Publisher site
See Article on Publisher Site

Abstract

For existing object detectors, anchor‐based detectors lack global information, while anchor‐free detectors based on key points lack prior position information. The above issues may lead to the imbalance of detection accuracy and shape robustness. In order to alleviate the above contradictions, a new auxiliary network key point generation network (GPNet) is proposed to improve the performance of existing object detectors. Specifically, a series of key points are generated through ground truth (GT) supervision to obtain more global information. These key points generate pseudo boxes (Pbox) with learnable parameters. Pbox has more specific prior information than the manually designed candidate boxes. The scale information of the Pbox is embed into the classification branch to obtain a more appropriate receptive field. In addition, a novel improved strategy for label assignment by combining Pbox and GT to enhance the ability to classify positive and negative samples is proposed. Extensive experiments on multiple dense prediction methods validate the effectiveness of GPNet, with a performance improvement of 1.5 AP over baseline. In particular, with ResNext‐101‐64× 4d‐DCN as the backbone, this method achieves 49.5 AP with single‐scale testing.

Journal

Advanced Theory and Simulations – Wiley

Published: May 1, 2023

Keywords: deep convolutional networks; label assignment; object detection; pyramid deformable convolution

References

  • Proc. of the IEEE Conf. on Computer Vision and Pattern Recognition
    Long, J.; Shelhamer, E.; Darrell, T.
  • Proc. of the IEEE Conf. on Computer Vision and Pattern Recognition
    Redmon, J.; Divvala, S.; Girshick, R.; Farhadi, A.
  • Proc. of the IEEE Int. Conf. on Computer Vision
    Lin, T.‐Y.; Goyal, P.; Girshick, R.; He, K.; Dollár, P.
  • Proc. of the IEEE/CVF International Conference on Computer Vision
    Tian, Z.; Shen, C.; Chen, H.; He, T.
  • Proc. of the IEEE/CVF Conf. on Computer Vision and Pattern Recognition
    Zhang, S.; Chi, C.; Yao, Y.; Lei, Z.; Li, S. Z.
  • Proc. of the IEEE Conf. on Computer Vision and Pattern Recognition
    Chen, X.; Kundu, K.; Zhang, Z.; Ma, H.; Fidler, S.; Urtasun, R.
  • Proc. of the IEEE Int. Conf. on Computer Vision
    Rad, M.; Lepetit, V.
  • Proc. of the IEEE/CVF Conf. on Computer Vision and Pattern Recognition
    Chen, Y.; Cao, Y.; Hu, H.; Wang, L.
  • Proc. of the IEEE Conf. on Computer Vision and Pattern Recognition
    Cai, Z.; Vasconcelos, N.
  • European Conf. on Computer Vision
    Liu, W.; Anguelov, D.; Erhan, D.; Szegedy, C.; Reed, S.; Fu, C.‐Y.; Berg, A. C.
  • Proc. of the IEEE/CVF Int. Conf. on Computer Vision
    Yang, Z.; Liu, S.; Hu, H.; Wang, L.; Lin, S.
  • Proc. of the European Conf. on Computer Vision (ECCV)
    Law, H.; Deng, J.
  • Proc. of the IEEE/CVF Int. Conf. on Computer Vision
    Duan, K.; Bai, S.; Xie, L.; Qi, H.; Huang, Q.; Tian, Q.
  • Proc. of the IEEE Conf. on Computer Vision and Pattern Recognition
    Xie, S.; Girshick, R.; Dollár, P.; Tu, Z.; He, K.
  • Proc. of the IEEE Int. Conf. on Computer Vision
    He, K.; Gkioxari, G.; Dollár, P.; Girshick, R.
  • Proc. of the IEEE/CVF Conf. on Computer Vision and Pattern Recognition
    Li, X.; Wang, W.; Hu, X.; Li, J.; Tang, J.; Yang, J.
  • Proc. of the IEEE/CVF Conf. on Computer Vision and Pattern Recognition
    Zhou, X.; Zhuo, J.; Krahenbuhl, P.
  • European Conf. on Computer Vision
    Kim, K.; Lee, H. S.
  • Proc. of the IEEE Int. Conf. on Computer Vision (ICCV)
    Girshick, R.
  • Proc. of the IEEE/CVF Conf. on Computer Vision and Pattern Recognition
    Rezatofighi, H.; Tsoi, N.; Gwak, J.; Sadeghian, A.; Reid, I.; Savarese, S.
  • European Conf. on Computer Vision
    Lin, T.‐Y.; Maire, M.; Belongie, S.; Hays, J.; Perona, P.; Ramanan, D.; Dollár, P.; Zitnick, C. L.
  • Proc. of the IEEE Conf. on Computer Vision and Pattern Recognition
    He, K.; Zhang, X.; Ren, S.; Sun, J.
  • Proc. of IEEE Conf. Computer Vision and Pattern Recognition
    Deng, J.; Dong, W.; Socher, R.; Li, L. ‐J.; Li, K.; Fei‐Fei, L.
  • 17th European Conf. on Computer Vision
    Zand, M.; Etemad, A.; Greenspan, M.
  • Proc. of the IEEE/CVF Conf. on Computer Vision and Pattern Recognition
    Li, S.; He, C.; Li, R.; Zhang, L.