DPPD: Deformable Polar Polygon Object Detection

被引:3
|
作者
Zheng, Yang [1 ]
Andrienko, Oles [1 ]
Zhao, Yonglei [1 ]
Park, Minwoo [1 ]
Pham, Trung [1 ]
机构
[1] NVIDIA, Santa Clara, CA 95051 USA
来源
2023 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION WORKSHOPS, CVPRW | 2023年
关键词
D O I
10.1109/CVPRW59228.2023.00013
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Regular object detection methods output rectangle bounding boxes, which are unable to accurately describe the actual object shapes. Instance segmentation methods output pixel-level labels, which are computationally expensive for real-time applications. Therefore, a polygon representation is needed to achieve precise shape alignment, while retaining low computation cost. We develop a novel Deformable Polar Polygon Object Detection method (DPPD) to detect objects in polygon shapes. In particular, our network predicts, for each object, a sparse set of flexible vertices to construct the polygon, where each vertex is represented by a pair of angle and distance in the Polar coordinate system. To enable training, both ground truth and predicted polygons are densely resampled to have the same number of vertices with equal-spaced raypoints. The resampling operation is fully differentable, allowing gradient back-propagation. Sparse polygon predicton ensures high-speed runtime inference while dense resampling allows the network to learn object shapes with high precision. The polygon detection head is established on top of an anchor-free and NMS-free network architecture. DPPD has been demonstrated successfully in various object detection tasks for autonomous driving such as traffic-sign, crosswalk, vehicle and pedestrian objects.
引用
收藏
页码:78 / 87
页数:10
相关论文
共 50 条
  • [1] Polygon-to-Polygon Distance Loss for Rotated Object Detection
    Yang, Yang
    Chen, Jifeng
    Zhong, Xiaopin
    Deng, Yuanlong
    THIRTY-SIXTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE / THIRTY-FOURTH CONFERENCE ON INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE / THE TWELVETH SYMPOSIUM ON EDUCATIONAL ADVANCES IN ARTIFICIAL INTELLIGENCE, 2022, : 3072 - 3080
  • [2] Deformable Capsules for Object Detection
    LaLonde, Rodney
    Khosravan, Naji
    Bagci, Ulas
    ADVANCED INTELLIGENT SYSTEMS, 2024, 6 (09)
  • [3] GJK for deformable object collision detection
    Hatab, Maher
    Kheddar, Abderrahmane
    2006 IEEE INTERNATIONAL WORKSHOP ON HAPTIC AUDIO VISUAL ENVIRONMENTS AND THEIR APPLICATIONS, 2006, : 147 - +
  • [4] Object detection and recognition via deformable illumination and deformable shape
    Zhou, Qiang
    Ma, Limin
    Celenk, Mehmet
    Chelberg, David
    2006 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING, ICIP 2006, PROCEEDINGS, 2006, : 2737 - +
  • [5] Constellational contour parsing for deformable object detection
    Huang, Chen
    Han, Tony X.
    He, Zhihai
    Cao, Wenming
    JOURNAL OF VISUAL COMMUNICATION AND IMAGE REPRESENTATION, 2016, 38 : 540 - 549
  • [6] Shape Band: A Deformable Object Detection Approach
    Bai, Xiang
    Li, Quannan
    Latecki, Longin Jan
    Liu, Wenyu
    Tu, Zhuowen
    CVPR: 2009 IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, VOLS 1-4, 2009, : 1335 - +
  • [7] OBJECT DETECTION VIA BOOSTED DEFORMABLE FEATURES
    Hussein, Mohamed
    Porikli, Fatih
    Davis, Larry
    2009 16TH IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING, VOLS 1-6, 2009, : 1445 - +
  • [8] Deformable Template Network (DTN) for Object Detection
    Wu, Shuai
    Xu, Yong
    Zhang, Bob
    Yang, Jian
    Zhang, David
    IEEE Transactions on Multimedia, 2022, 24 : 2058 - 2068
  • [9] Object Detection Based on Deformable Part Model
    Wei Lei
    Xu Zhiyong
    8TH INTERNATIONAL SYMPOSIUM ON ADVANCED OPTICAL MANUFACTURING AND TESTING TECHNOLOGY: OPTICAL TEST, MEASUREMENT TECHNOLOGY, AND EQUIPMENT, 2016, 9684
  • [10] Visual Object Detection with Deformable Part Models
    Felzenszwalb, Pedro
    Girshick, Ross
    McAllester, David
    Ramanan, Deva
    COMMUNICATIONS OF THE ACM, 2013, 56 (09) : 97 - 105