Adaptive Object Detection Using Adjacency and Zoom Prediction

被引:37
|
作者
Lu, Yongxi [1 ]
Javidi, Tara [1 ]
Lazebnik, Svetlana [2 ]
机构
[1] Univ Calif San Diego, San Diego, CA 92103 USA
[2] Univ Illinois, Urbana, IL 61801 USA
基金
美国国家科学基金会;
关键词
D O I
10.1109/CVPR.2016.258
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
State-of-the-art object detection systems rely on an accurate set of region proposals. Several recent methods use a neural network architecture to hypothesize promising object locations. While these approaches are computationally efficient, they rely on fixed image regions as anchors for predictions. In this paper we propose to use a search strategy that adaptively directs computational resources to sub-regions likely to contain objects. Compared to methods based on fixed anchor locations, our approach naturally adapts to cases where object instances are sparse and small. Our approach is comparable in terms of accuracy to the state-of-the-art Faster R-CNN approach while using two orders of magnitude fewer anchors on average. Code is publicly available.
引用
收藏
页码:2351 / 2359
页数:9
相关论文
共 50 条
  • [1] Color object detection using pyramidal adjacency graphs
    Lozano, V
    Colantoni, P
    Laget, B
    [J]. INTERNATIONAL CONFERENCE ON IMAGE PROCESSING, PROCEEDINGS - VOL III, 1996, : 1007 - 1010
  • [2] REGION ADJACENCY AND ITS APPLICATION TO OBJECT DETECTION
    SUK, M
    OH, SJ
    [J]. PATTERN RECOGNITION, 1986, 19 (02) : 161 - 167
  • [3] Adaptive Imaging for Lesion Detection Using a Zoom-in PET System
    Zhou, Jian
    Qi, Jinyi
    [J]. IEEE TRANSACTIONS ON MEDICAL IMAGING, 2011, 30 (01) : 119 - 130
  • [4] AANet: Adjacency auxiliary network for salient object detection
    Li, Xialu
    Cui, Ziguan
    Gan, Zongliang
    Tang, Guijin
    Liu, Feng
    [J]. KSII TRANSACTIONS ON INTERNET AND INFORMATION SYSTEMS, 2021, 15 (10): : 3729 - 3749
  • [5] Towards Efficient Multiview Object Detection with Adaptive Action Prediction
    Xu, Qianli
    Fang, Fen
    Gauthier, Nicolas
    Liang, Wenyu
    Wu, Yan
    Li, Liyuan
    Lim, Joo-Hwee
    [J]. 2021 IEEE INTERNATIONAL CONFERENCE ON ROBOTICS AND AUTOMATION (ICRA 2021), 2021, : 13423 - 13429
  • [6] ADJACENCY DETECTION USING QUADCODES
    LI, SX
    LOEW, MH
    [J]. COMMUNICATIONS OF THE ACM, 1987, 30 (07) : 627 - 631
  • [7] Object detection based on weighted adaptive prediction in lifting scheme transform
    Amiri, Mahdi
    Rabiee, Hamid R.
    [J]. ISM 2006: EIGHTH IEEE INTERNATIONAL SYMPOSIUM ON MULTIMEDIA, PROCEEDINGS, 2006, : 652 - +
  • [8] Small Object Detection using Prediction head and Attention
    Kim, Hae Moon
    Kim, Ji Hoon
    Park, Kyung Ri
    Moon, Young Shik
    [J]. 2022 INTERNATIONAL CONFERENCE ON ELECTRONICS, INFORMATION, AND COMMUNICATION (ICEIC), 2022,
  • [9] Multiple Moving Object Detection From UAV Videos Using Trajectories of Matched Regional Adjacency Graphs
    Kalantar, Bahareh
    Bin Mansor, Shattri
    Halin, Alfian Abdul
    Shafri, Helmi Zulhaidi Mohd
    Zand, Mohsen
    [J]. IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2017, 55 (09): : 5198 - 5213
  • [10] AN OBJECT-DETECTION ALGORITHM BASED ON THE REGION-ADJACENCY GRAPH
    SUK, M
    CHO, TH
    [J]. PROCEEDINGS OF THE IEEE, 1984, 72 (07) : 985 - 986