The Object Detection of Anchor Free—A Survey
Rui Huo
School of Physics and Electronic Information, Yan’an University
Shuili Zhang
School of Physics and Electronic Information, Yan’an University
Huaiyuan Sun
School of Physics and Electronic Information, Yan’an University
Xve Tian
School of Physics and Electronic Information, Yan’an University
DOI: https://doi.org/10.59429/pmcs.v6i3.7244
Keywords: Deep learning; Object detection; Image recognition; Anchor
Abstract
Since Ross Girshick introduced deep learning to object detection and proposed RCNN networks in 2014, the field of object detection has been blasting ahead. Deep learning-based object detection algorithms typically classify and regress region proposals. In the one-stage detector, these region proposals are the anchor boxes generated by the sliding window approach. In this paper, we first give a brief overview of anchors and compare the advantages and disadvantages of both anchor base and anchor-free object detection. Then, we collect and organize the anchor-free object detection algorithms (such as CenterNet, FCOS, FSAF, etc.) that have received much attention and use in recent years. We also present a detailed description of the anchor-free object detection algorithms in three parts (keypoint-based, Segmentation-based, and YOLO series).
References
[1] Ren S Q, He K M, Girshick R,Sun J (2017) Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks. IEEE Transactions on Pattern Analysis and Machine Intelligence 39:1137-1149
[2] Redmon J, Divvala S, Girshick R, Farhadi A,Ieee (2016) You Only Look Once: Unified, Real-Time Object Detection. In: 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR),PP 779-788
[3] Li B, Shi Y, Qi Z Q,Chen Z S (2018) A Survey on Semantic Segmentation. In: 18th IEEE International Conference on Data Mining Workshops (ICDMW),PP 1233-1240 [4] Duan K W, Bai S, Xie L X, Qi H G, Huang Q M, Tian Q,Ieee (2019) CenterNet: Keypoint Triplets for Object Detection. In: IEEE/CVF International Conference on Computer Vision (ICCV),PP 6568-6577
[5] Zhou X, Wang D,Krhenbühl P (2019) Objects as Points. arXiv:1904.07850
[6] He K, Zhang X, Ren S,Sun J (2016) Deep Residual Learning for Image Recognition. In: 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR),PP 770-778
[7] Yu F, Wang D, Shelhamer E,Darrell T (2018) Deep Layer Aggregation. In: 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition,PP 2403-2412
[8] Zhu C C, He Y H, Savvides M,Soc I C (2019) Feature Selective Anchor-Free Module for SingleShot Object Detection. In: 32nd IEEE/CVF Conference on ComputerVision and Pattern Recognition (CVPR),PP 840-849
[9] Lin T Y, Goyal P, Girshick R, He K,Dollár P (2020) Focal Loss for Dense Object Detection. IEEE Transactions on Pattern Analysis and Machine Intelligence 42:318-327
[10] Tian Z, Shen C, Chen H,He T (2019) FCOS: Fully Convolutional One-Stage Object Detection. In: 2019 IEEE/CVF International Conference on Computer Vision (ICCV),PP 9626-9635
[11] Qin X, Zhang Z, Huang C, Gao C, Dehghan M,Jagersand M (2019) BASNet: Boundary-Aware Salient Object Detection. In: 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR),PP 7471-7481
[12] He K, Gkioxari G, Dollar P,Girshick R (2017) Mask R-CNN. arXiv:1703.06870v3
[13] He X, Zhao K,Chu X (2019) AutoML: A Survey of the State-of-the-Art. arXiv:1908.00709
[14] Zoph B,Le Q V (2016) Neural Architecture Search with Reinforcement Learning. arXiv:1611.01578
[15] Kong T, Sun F, Liu H, Jiang Y, Li L,Shi J (2020) FoveaBox: Beyound Anchor-Based Object Detection. IEEE Transactions on Image Processing 29:7389-7398
[16] Iandola F N, Han S, Moskewicz M W, Ashraf K, Dally W J,Keutzer K (2016) SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size. arXiv:1602.07360
[17] Zhu C, Chen F, Shen Z,Savvides M (2019) Soft Anchor-Point Object Detection. arXiv:1911.12448