SelaVPR++: Towards Seamless Adaptation of Foundation Models for Efficient Place Recognition
IEEE Transactions on Pattern Analysis and Machine Intelligence · 2026-03 · 48 (3) · 2731-2748
DOI: 10.1109/tpami.2025.3629287 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Pattern Analysis and Machine Intelligence, 2026-03. Verfasst von Feng Lu und 6 weiteren Autor:innen. Die Arbeit wurde bisher 10-mal zitiert (Crossref) und nennt 82 Literaturangaben. Gefördert von: National Key R&D Program of China, SSTIC, Key Project of Pengcheng Laboratory.
Abstract (Originalsprache)
Recent studies show that the visual place recognition (VPR) method using pre-trained visual foundation models can achieve promising performance. In our previous work, we propose a novel method to realize seamless adaptation of foundation models to VPR (SelaVPR). This method can produce both global and local features that focus on discriminative landmarks to recognize places for two-stage VPR by a parameter-efficient adaptation approach. Although SelaVPR has achieved competitive results, we argue that the previous adaptation is inefficient in training time and GPU memory usage, and the re-ranking paradigm is also costly in retrieval latency and storage usage. In pursuit of higher efficiency and better performance, we propose an extension of the SelaVPR, called SelaVPR++. Concretely, we first design a parameter-, time-, and memory-efficient adaptation method that uses lightweight multi-scale convolution (MultiConv) adapters to refine intermediate features from the frozen foundation backbone. This adaptation method does not back-propagate gradients through the backbone during training, and the MultiConv adapter facilitates feature interactions along the spatial axes and introduces proper local priors, thus achieving higher efficiency and better performance. Moreover, we propose an innovative re-ranking paradigm for more efficient VPR. Instead of relying on local features for re-ranking, which incurs huge overhead in latency and storage, we employ compact binary features for initial retrieval and robust floating-point (global) features for re-ranking. To obtain such binary features, we propose a similarity-constrained deep hashing method, which can be easily integrated into the VPR pipeline. Finally, we improve our training strategy and unify the training protocol of several common training datasets to merge them for better training of VPR models. Extensive experiments show that SelaVPR++ is highly efficient in training time, GPU memory usage, and retrieval latency (6000× faster than TransVPR), as well as outperforms the state-of-the-art methods by a large margin (ranks 1st on MSLS challenge leaderboard).
Förderung
- National Key R&D Program of China
- SSTIC
- Key Project of Pengcheng Laboratory
- National Natural Science Foundation of China
Zitieren
APA
Lu, F., Jin, T., Lan, X., Zhang, L., Liu, Y., Wang, Y., & Yuan, C. (2026). SelaVPR++: Towards Seamless Adaptation of Foundation Models for Efficient Place Recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 48(3), 2731-2748. https://doi.org/10.1109/tpami.2025.3629287
Vancouver
Lu F, Jin T, Lan X, Zhang L, Liu Y, Wang Y, et al. SelaVPR++: Towards Seamless Adaptation of Foundation Models for Efficient Place Recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2026;48(3):2731-2748. doi:10.1109/tpami.2025.3629287
BibTeX
@article{Lu2026,
author = {Lu, Feng and Jin, Tong and Lan, Xiangyuan and Zhang, Lijun and Liu, Yunpeng and Wang, Yaowei and Yuan, Chun},
title = {SelaVPR++: Towards Seamless Adaptation of Foundation Models for Efficient Place Recognition},
journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
year = {2026},
volume = {48},
number = {3},
pages = {2731--2748},
doi = {10.1109/tpami.2025.3629287}
}
EUVIMED