# Global Proxy-based Hard Mining for Visual Place Recognition

arXiv: https://arxiv.org/abs/2302.14217 · HTML: https://mlsanity.com/paper/2302.14217/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- Builds on 17 of its 37 references that are in the map.
- First posted on arXiv 28 Feb 2023.

## Builds on (most cited first, 17 of 17)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [FaceNet: A Unified Embedding for Face Recognition and Clustering](https://mlsanity.com/paper/1503.03832/) (2015)
- [NetVLAD: CNN architecture for weakly supervised place recognition](https://mlsanity.com/paper/1511.07247/) (2015)
- [In Defense of the Triplet Loss for Person Re-Identification](https://mlsanity.com/paper/1703.07737/) (2017)
- [Multi-Similarity Loss with General Pair Weighting for Deep Metric Learning](https://mlsanity.com/paper/1904.06627/) (2019)
- [A Metric Learning Reality Check](https://mlsanity.com/paper/2003.08505/) (2020)
- [Sampling Matters in Deep Embedding Learning](https://mlsanity.com/paper/1706.07567/) (2017)
- [Patch-NetVLAD: Multi-Scale Fusion of Locally-Global Descriptors for Place Recognition](https://mlsanity.com/paper/2103.01486/) (2021)
- [Proxy Anchor Loss for Deep Metric Learning](https://mlsanity.com/paper/2003.13911/) (2020)
- [Efficient Deep Learning: A Survey on Making Deep Learning Models Smaller, Faster, and Better](https://mlsanity.com/paper/2106.08962/) (2021)
- [Deep Metric Learning with Hierarchical Triplet Loss](https://mlsanity.com/paper/1810.06951/) (2018)
- [TransVPR: Transformer-based place recognition with multi-level attention aggregation](https://mlsanity.com/paper/2201.02001/) (2022)
- [VPR-Bench: An Open-Source Visual Place Recognition Evaluation Framework with Quantifiable Viewpoint and Appearance Change](https://mlsanity.com/paper/2005.08135/) (2020)
- [GSV-Cities: Toward Appropriate Supervised Visual Place Recognition](https://mlsanity.com/paper/2210.10239/) (2022)
- [k-Nearest Neighbour Classifiers: 2nd Edition (with Python examples)](https://mlsanity.com/paper/2004.04523/) (2020)
- [Hierarchical Proxy-based Loss for Deep Metric Learning](https://mlsanity.com/paper/2103.13538/) (2021)
- [Domain Adaptation of Learned Features for Visual Localization](https://mlsanity.com/paper/2008.09310/) (2020)

## Built on by (most cited first, 8 of 8)

- [BoQ: A Place is Worth a Bag of Learnable Queries](https://mlsanity.com/paper/2405.07364/) (2024)
- [VG-SSL: Benchmarking Self-supervised Representation Learning Approaches for Visual Geo-localization](https://mlsanity.com/paper/2308.00090/) (2023)
- [ClusVPR: Efficient Visual Place Recognition with Clustering-based Weighted Transformer](https://mlsanity.com/paper/2310.04099/) (2023)
- [SAGE: Spatial-visual Adaptive Graph Exploration for Efficient Visual Place Recognition](https://mlsanity.com/paper/2509.25723/) (2025)
- [AstroLoc: Robust Space to Ground Image Localizer](https://mlsanity.com/paper/2502.07003/) (2025)
- [DialogueVPR: Towards Conversational Visual Place Recognition](https://mlsanity.com/paper/2607.14115/) (2026)
- [One Channel to Rule Them All: Rethinking Input Representation for Visual Place Recognition](https://mlsanity.com/paper/2606.00936/) (2026)
- [AdaptVPR: Route-Aware Hard Positive Generation for Robust Visual Place Recognition](https://mlsanity.com/paper/2609.04369/) (2026)

## Abstract (arXiv)

Learning deep representations for visual place recognition is commonly performed using pairwise or triple loss functions that highly depend on the hardness of the examples sampled at each training iteration. Existing techniques address this by using computationally and memory expensive offline hard mining, which consists of identifying, at each iteration, the hardest samples from the training set. In this paper we introduce a new technique that performs global hard mini-batch sampling based on proxies. To do so, we add a new end-to-end trainable branch to the network, which generates efficient place descriptors (one proxy for each place). These proxy representations are thus used to construct a global index that encompasses the similarities between all places in the dataset, allowing for highly informative mini-batch sampling at each training iteration. Our method can be used in combination with all existing pairwise and triplet loss functions with negligible additional memory and computation cost. We run extensive ablation studies and show that our technique brings new state-of-the-art performance on multiple large-scale benchmarks such as Pittsburgh, Mapillary-SLS and SPED. In particular, our method provides more than 100% relative improvement on the challenging Nordland dataset. Our code is available at https://github.com/amaralibey/GPM