Joint Multi-Condition Representation Modelling via Matrix Factorisation for Visual Place Recognition
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908603667972096 |
|---|---|
| author | Ismagilov, Timur Majeed, Shakaiba Milford, Michael Nguyen, Tan Viet Tuyen Ramchurn, Sarvapali D. Ehsan, Shoaib |
| author_facet | Ismagilov, Timur Majeed, Shakaiba Milford, Michael Nguyen, Tan Viet Tuyen Ramchurn, Sarvapali D. Ehsan, Shoaib |
| contents | We address multi-reference visual place recognition (VPR), where reference sets captured under varying conditions are used to improve localisation performance. While deep learning with large-scale training improves robustness, increasing data diversity and model complexity incur extensive computational cost during training and deployment. Descriptor-level fusion via voting or aggregation avoids training, but often targets multi-sensor setups or relies on heuristics with limited gains under appearance and viewpoint change. We propose a training-free, descriptor-agnostic approach that jointly models places using multiple reference descriptors via matrix decomposition into basis representations, enabling projection-based residual matching. We also introduce SotonMV, a structured benchmark for multi-viewpoint VPR. On multi-appearance data, our method improves Recall@1 by up to ~18% over single-reference and outperforms multi-reference baselines across appearance and viewpoint changes, with gains of ~5% on unstructured data, demonstrating strong generalisation while remaining lightweight. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_17739 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Joint Multi-Condition Representation Modelling via Matrix Factorisation for Visual Place Recognition Ismagilov, Timur Majeed, Shakaiba Milford, Michael Nguyen, Tan Viet Tuyen Ramchurn, Sarvapali D. Ehsan, Shoaib Computer Vision and Pattern Recognition We address multi-reference visual place recognition (VPR), where reference sets captured under varying conditions are used to improve localisation performance. While deep learning with large-scale training improves robustness, increasing data diversity and model complexity incur extensive computational cost during training and deployment. Descriptor-level fusion via voting or aggregation avoids training, but often targets multi-sensor setups or relies on heuristics with limited gains under appearance and viewpoint change. We propose a training-free, descriptor-agnostic approach that jointly models places using multiple reference descriptors via matrix decomposition into basis representations, enabling projection-based residual matching. We also introduce SotonMV, a structured benchmark for multi-viewpoint VPR. On multi-appearance data, our method improves Recall@1 by up to ~18% over single-reference and outperforms multi-reference baselines across appearance and viewpoint changes, with gains of ~5% on unstructured data, demonstrating strong generalisation while remaining lightweight. |
| title | Joint Multi-Condition Representation Modelling via Matrix Factorisation for Visual Place Recognition |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2510.17739 |