Learning to Optimize Non-Rigid Tracking

被引:15
|
作者
Li, Yang [1 ,4 ]
Bozic, Aljaz [4 ]
Zhang, Tianwei [1 ]
Ji, Yanli [1 ,2 ]
Harada, Tatsuya [1 ,3 ]
Niessner, Matthias [4 ]
机构
[1] Univ Tokyo, Tokyo, Japan
[2] UESTC, Chengdu, Peoples R China
[3] RIKEN, Tokyo, Japan
[4] Tech Univ Munich, Munich, Germany
关键词
D O I
10.1109/CVPR42600.2020.00496
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
One of the widespread solutions for non-rigid tracking has a nested-loop structure: with Gauss-Newton to minimize a tracking objective in the outer loop, and Preconditioned Conjugate Gradient (PCG) to solve a sparse linear system in the inner loop. In this paper, we employ learnable optimizations to improve tracking robustness and speed up solver convergence. First, we upgrade the tracking objective by integrating an alignment data term on deep features which are learned end-to-end through CNN. The new tracking objective can capture the global deformation which helps Gauss-Newton to jump over local minimum, leading to robust tracking on large non-rigid motions. Second, we bridge the gap between the preconditioning technique and learning method by introducing a ConditionNet which is trained to generate a preconditioner such that PCG can converge within a small number of steps. Experimental results indicate that the proposed learning method converges faster than the original PCG by a large margin.
引用
收藏
页码:4909 / 4917
页数:9
相关论文
共 50 条
  • [1] Non-rigid Object Tracking
    Zhou, Huiyu
    Schaefer, Gerald
    [J]. PROCEEDINGS ELMAR-2010, 2010, : 101 - 104
  • [2] Neural Non-Rigid Tracking
    Bozic, Aljaz
    Palafox, Pablo
    Zollhoefer, Michael
    Dai, Angela
    Thies, Justus
    Niessner, Matthias
    [J]. ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 33, NEURIPS 2020, 2020, 33
  • [3] Learning-based tracking of complex non-rigid motion
    Qiang Wang
    Hai-Zhou Ai
    Guang-You Xu
    [J]. Journal of Computer Science and Technology, 2004, 19 : 489 - 500
  • [4] Learning-based tracking of complex non-rigid motion
    Wang, Q
    Ai, HZ
    Xu, GY
    [J]. JOURNAL OF COMPUTER SCIENCE AND TECHNOLOGY, 2004, 19 (04) : 489 - 500
  • [5] Deep Learning Based Non-rigid Device Tracking in Ultrasound Image
    Chen, Shaohan
    Wang, Shu
    [J]. PROCEEDINGS OF 2018 THE 2ND INTERNATIONAL CONFERENCE ON COMPUTER SCIENCE AND ARTIFICIAL INTELLIGENCE (CSAI 2018) / 2018 THE 10TH INTERNATIONAL CONFERENCE ON INFORMATION AND MULTIMEDIA TECHNOLOGY (ICIMT 2018), 2018, : 354 - 358
  • [6] TRACKING NON-RIGID STRUCTURES IN COMPUTER SIMULATIONS
    Gezahegne, Abel
    Kainath, Chandrika
    [J]. 2008 15TH IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING, VOLS 1-5, 2008, : 1548 - 1551
  • [7] Towards Optimal Non-rigid Surface Tracking
    Klaudiny, Martin
    Budd, Chris
    Hilton, Adrian
    [J]. COMPUTER VISION - ECCV 2012, PT IV, 2012, 7575 : 743 - 756
  • [8] Non-rigid object tracking in complex scenes
    Zhou, Huiyu
    Yuan, Yuan
    Zhang, Yi
    Shi, Chunmei
    [J]. PATTERN RECOGNITION LETTERS, 2009, 30 (02) : 98 - 102
  • [9] A new algorithm to rigid and non-rigid object tracking in complex environments
    Electrical Engineering Department, Islamic Azad University , South Tehran Branch, Tehran, Iran
    [J]. Int J Adv Manuf Technol, 2013, 9-12 (1643-1651):
  • [10] A new algorithm to rigid and non-rigid object tracking in complex environments
    Mazinan, A. H.
    Amir-Latifi, A.
    [J]. INTERNATIONAL JOURNAL OF ADVANCED MANUFACTURING TECHNOLOGY, 2013, 64 (9-12): : 1643 - 1651