Print Email Facebook Twitter ConvSequential-SLAM Title ConvSequential-SLAM: A Sequence-Based, Training-Less Visual Place Recognition Technique for Changing Environments Author Tomia, Mihnea Alexandru (University of Essex) Zaffar, M. (TU Delft Intelligent Vehicles) Milford, Michael J. (Queensland University of Technology) McDonald-Maier, Klaus D. (University of Essex) Ehsan, Shoaib (University of Essex) Date 2021 Abstract Visual Place Recognition (VPR) is the ability to correctly recall a previously visited place under changing viewpoints and appearances. A large number of handcrafted and deep-learning-based VPR techniques exist, where the former suffer from appearance changes and the latter have significant computational needs. In this paper, we present a new handcrafted VPR technique, namely ConvSequential-SLAM, that achieves state-of-the-art place matching performance under challenging conditions. We utilise sequential information and block-normalisation to handle appearance changes, while using regional-convolutional matching to achieve viewpoint-invariance. We analyse content-overlap in-between query frames to find a minimum sequence length, while also re-using the image entropy information for environment-based sequence length tuning. State-of-the-art performance is reported in contrast to 9 contemporary VPR techniques on 4 public datasets. Qualitative insights and an ablation study on sequence length are also provided. Subject sequence-based filteringSLAMvisual localizationvisual place recognition To reference this document use: http://resolver.tudelft.nl/uuid:81dd8472-c864-425e-a064-370715ad3a98 DOI https://doi.org/10.1109/ACCESS.2021.3107778 ISSN 2169-3536 Source IEEE Access, 9, 118673-118683 Part of collection Institutional Repository Document type journal article Rights © 2021 Mihnea Alexandru Tomia, M. Zaffar, Michael J. Milford, Klaus D. McDonald-Maier, Shoaib Ehsan Files PDF ConvSequential_SLAM_A_Seq ... nments.pdf 2 MB Close viewer /islandora/object/uuid:81dd8472-c864-425e-a064-370715ad3a98/datastream/OBJ/view