PBDE: an effective post-processing method based on box density for object detection

An inevitable of object detection is the existence of false positive detection boxes. The existence of a large number of false detection boxes greatly reduces the precision and compromises the desired effect. In this paper, we propose a post-processing method named Prediction Box Density Evaluation (PBDE). During applying object detect models to actual application scenarios, we summarize box density characteristics of true positive (TP) and false positive (FP) boxes. Then we set a threshold of box density to filter out a large number of FP boxes. After applying the PBDE algorithm, we obtain a significant improvement in precision and F1-Score. We have verified the effectiveness of our post-processing method in different application scenarios and models. The entire algorithm is carried out on the post-processing of object detection. There is no need to change the original training method and network structure, which is of great practicality and generality.

[1]  Ali Farhadi,et al.  YOLOv3: An Incremental Improvement , 2018, ArXiv.

[2]  Kaiming He,et al.  Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks , 2015, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[3]  Pietro Perona,et al.  Microsoft COCO: Common Objects in Context , 2014, ECCV.

[4]  Larry S. Davis,et al.  Soft-NMS — Improving Object Detection with One Line of Code , 2017, 2017 IEEE International Conference on Computer Vision (ICCV).

[5]  Qi Tian,et al.  CenterNet: Keypoint Triplets for Object Detection , 2019, 2019 IEEE/CVF International Conference on Computer Vision (ICCV).

[6]  Zhe Cui,et al.  Light-YOLOv3: fast method for detecting green mangoes in complex scenes using picking robots , 2020, Applied Intelligence.

[7]  Hyunchul Shin,et al.  Context-aware pedestrian detection especially for small-sized instances with Deconvolution Integrated Faster RCNN (DIF R-CNN) , 2018, Applied Intelligence.

[8]  Mark Sandler,et al.  MobileNetV2: Inverted Residuals and Linear Bottlenecks , 2018, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[9]  Christopher G. Harris,et al.  A Combined Corner and Edge Detector , 1988, Alvey Vision Conference.

[10]  Luc Van Gool,et al.  A mobile vision system for robust multi-person tracking , 2008, 2008 IEEE Conference on Computer Vision and Pattern Recognition.

[11]  Xiangyu Zhang,et al.  Softer-NMS: Rethinking Bounding Box Regression for Accurate Object Detection , 2018, ArXiv.

[12]  Quoc V. Le,et al.  Searching for MobileNetV3 , 2019, 2019 IEEE/CVF International Conference on Computer Vision (ICCV).

[13]  Hong-Yuan Mark Liao,et al.  YOLOv4: Optimal Speed and Accuracy of Object Detection , 2020, ArXiv.

[14]  David G. Lowe,et al.  Distinctive Image Features from Scale-Invariant Keypoints , 2004, International Journal of Computer Vision.

[15]  Ross B. Girshick,et al.  Fast R-CNN , 2015, 1504.08083.

[16]  Bo Chen,et al.  MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications , 2017, ArXiv.

[17]  Yuxing Peng,et al.  ThunderNet: Towards Real-Time Generic Object Detection on Mobile Devices , 2019, 2019 IEEE/CVF International Conference on Computer Vision (ICCV).

[18]  Quoc V. Le,et al.  EfficientDet: Scalable and Efficient Object Detection , 2019, 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[19]  Wei Liu,et al.  SSD: Single Shot MultiBox Detector , 2015, ECCV.

[20]  Paul A. Viola,et al.  Rapid object detection using a boosted cascade of simple features , 2001, Proceedings of the 2001 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. CVPR 2001.

[21]  A. Rosenfeld,et al.  Edge and Curve Detection for Visual Scene Analysis , 1971, IEEE Transactions on Computers.

[22]  Nuno Vasconcelos,et al.  Cascade R-CNN: Delving Into High Quality Object Detection , 2017, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[23]  Andreas Geiger,et al.  Vision meets robotics: The KITTI dataset , 2013, Int. J. Robotics Res..

[24]  Kaiming He,et al.  Mask R-CNN , 2017, 2017 IEEE International Conference on Computer Vision (ICCV).

[25]  Lei Xie,et al.  An Effective Face Anti-Spoofing Method via Stereo Matching , 2021, IEEE Signal Processing Letters.

[26]  Ali Farhadi,et al.  You Only Look Once: Unified, Real-Time Object Detection , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[27]  Haiwen Tu,et al.  Research on Accurate Prediction of the Container Ship Resistance by RBFNN and Other Machine Learning Algorithms , 2021, Journal of Marine Science and Engineering.

[28]  Yunhong Wang,et al.  Adaptive NMS: Refining Pedestrian Detection in a Crowd , 2019, 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[29]  Forrest N. Iandola,et al.  SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <1MB model size , 2016, ArXiv.

[30]  Shuzhi Sam Ge,et al.  Small traffic sign detection from large image , 2019, Applied Intelligence.

[31]  Xiangyu Zhang,et al.  Bounding Box Regression With Uncertainty for Accurate Object Detection , 2018, 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[32]  Hongye Su,et al.  A compression pipeline for one-stage object detection model , 2021, J. Real Time Image Process..

[33]  Kaiming He,et al.  Focal Loss for Dense Object Detection , 2017, 2017 IEEE International Conference on Computer Vision (ICCV).

[34]  Liang He,et al.  Fast and accurate cable detection using CNN , 2020, Applied Intelligence.