TCL: Tightly Coupled Learning Strategy for Weakly Supervised Hierarchical Place Recognition
- 10 January 2022
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 7 (2) , 2684-2691
Abstract
Visual place recognition (VPR) is a key issue for robotics and autonomous systems. For the trade-off between time and performance, most of methods use the coarse-to-fine hierarchical architecture, which consists of retrieving top-N candidates using global features, and re-ranking top-N with local features. However, since the two types of features are usually processed independently, re-ranking may harm global retrieval, termed re-ranking confusion . Moreover, re-ranking is limited by global retrieval. In this paper, we propose a tightly coupled learning (TCL) strategy to train triplet models. Different from original triplet learning (OTL) strategy, it combines global and local descriptors for joint optimization. In addition, a bidirectional search dynamic time warping (BS-DTW) algorithm is also proposed to mine locally spatial information tailored to VPR in re-ranking. The experimental results on public benchmarks show that the models using TCL outperform the models using OTL, and TCL can be used as a general strategy to improve performance for weakly supervised ranking tasks. Further, our lightweight unified model is better than several state-of-the-art methods and has over an order of magnitude of computational efficiency to meet the real-time requirements of robots.Keywords
All Related Versions
Funding Information
- National Natural Science Foundation of China (62088102)
- National Natural Science Foundation of China (61773312)
This publication has 31 references indexed in Scilit:
- Visual Place Recognition: A SurveyIEEE Transactions on Robotics, 2015
- On the performance of ConvNet features for place recognition2015
- Scene Signatures: Localised and Point-less Features for Localisation2014
- Visual Place Recognition with Repetitive Structures2013
- The New College Vision and Laser Data SetThe International Journal of Robotics Research, 2009
- FAB-MAP: Probabilistic Localization and Mapping in the Space of AppearanceThe International Journal of Robotics Research, 2008
- Object retrieval with large vocabularies and fast spatial matching2007
- SURF: Speeded Up Robust Features2006
- Video Google: a text retrieval approach to object matching in videos2003
- Object recognition from local scale-invariant features1999