# Towards Implicit Aggregation: Robust Image Representation for Place Recognition in the Transformer Era

arXiv: https://arxiv.org/abs/2511.06024 · HTML: https://mlsanity.com/paper/2511.06024/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- Builds on 48 of its 74 references that are in the map.
- First posted on arXiv 8 Nov 2025.

## Builds on (most cited first, 30 of 48)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [DINOv2: Learning Robust Visual Features without Supervision](https://mlsanity.com/paper/2304.07193/) (2023)
- [Parameter-Efficient Transfer Learning for NLP](https://mlsanity.com/paper/1902.00751/) (2019)
- [Prefix-Tuning: Optimizing Continuous Prompts for Generation](https://mlsanity.com/paper/2101.00190/) (2021)
- [The Power of Scale for Parameter-Efficient Prompt Tuning](https://mlsanity.com/paper/2104.08691/) (2021)
- [Visual Prompt Tuning](https://mlsanity.com/paper/2203.12119/) (2022)
- [NetVLAD: CNN architecture for weakly supervised place recognition](https://mlsanity.com/paper/1511.07247/) (2015)
- [ORB-SLAM3: An Accurate Open-Source Library for Visual, Visual-Inertial and Multi-Map SLAM](https://mlsanity.com/paper/2007.11898/) (2020)
- [Vision Transformers Need Registers](https://mlsanity.com/paper/2309.16588/) (2023)
- [Do Vision Transformers See Like Convolutional Neural Networks?](https://mlsanity.com/paper/2108.08810/) (2021)
- [Fine-tuning CNN Image Retrieval with No Human Annotation](https://mlsanity.com/paper/1711.02512/) (2017)
- [Multi-Similarity Loss with General Pair Weighting for Deep Metric Learning](https://mlsanity.com/paper/1904.06627/) (2019)
- [Patch-NetVLAD: Multi-Scale Fusion of Locally-Global Descriptors for Place Recognition](https://mlsanity.com/paper/2103.01486/) (2021)
- [Recurrent Memory Transformer](https://mlsanity.com/paper/2207.06881/) (2022)
- [You Only Look at One Sequence: Rethinking Transformer in Vision through Object Detection](https://mlsanity.com/paper/2106.00666/) (2021)
- [Rethinking Visual Geo-localization for Large-Scale Applications](https://mlsanity.com/paper/2204.02287/) (2022)
- [AnyLoc: Towards Universal Visual Place Recognition](https://mlsanity.com/paper/2308.00688/) (2023)
- [MixVPR: Feature Mixing for Visual Place Recognition](https://mlsanity.com/paper/2303.02190/) (2023)
- [Optimal Transport Aggregation for Visual Place Recognition](https://mlsanity.com/paper/2311.15937/) (2023)
- [EigenPlaces: Training Viewpoint Robust Models for Visual Place Recognition](https://mlsanity.com/paper/2308.10832/) (2023)
- [Self-supervising Fine-grained Region Similarities for Large-scale Image Localization](https://mlsanity.com/paper/2006.03926/) (2020)
- [TransVPR: Transformer-based place recognition with multi-level attention aggregation](https://mlsanity.com/paper/2201.02001/) (2022)
- [Deep Learning Features at Scale for Visual Place Recognition](https://mlsanity.com/paper/1701.05105/) (2017)
- [GSV-Cities: Toward Appropriate Supervised Visual Place Recognition](https://mlsanity.com/paper/2210.10239/) (2022)
- [Deep Visual Geo-localization Benchmark](https://mlsanity.com/paper/2204.03444/) (2022)
- [Memory Transformer](https://mlsanity.com/paper/2006.11527/) (2020)
- [On the Performance of ConvNet Features for Place Recognition](https://mlsanity.com/paper/1501.04158/) (2015)

## Built on by (most cited first, 6 of 6)

- [Region Matters: Efficient and Reliable Region-Aware Visual Place Recognition](https://mlsanity.com/paper/2604.22390/) (2026)
- [QdaVPR: A novel query-based domain-agnostic model for visual place recognition](https://mlsanity.com/paper/2603.07414/) (2026)
- [EPRBench: A High-Quality Benchmark Dataset for Event Stream Based Visual Place Recognition](https://mlsanity.com/paper/2602.12919/) (2026)
- [Are All Tokens Necessary for Visual Place Recognition? An Empirical Study of Token Reduction for Efficient Inference](https://mlsanity.com/paper/2607.15563/) (2026)
- [POI-Loc: A Fine-Grained POI Localization Benchmark and an Asymmetric Global-to-Local Matching Method](https://mlsanity.com/paper/2609.02012/) (2026)
- [AdaptVPR: Route-Aware Hard Positive Generation for Robust Visual Place Recognition](https://mlsanity.com/paper/2609.04369/) (2026)

## Abstract (arXiv)

Visual place recognition (VPR) is typically regarded as a specific image retrieval task, whose core lies in representing images as global descriptors. Over the past decade, dominant VPR methods (e.g., NetVLAD) have followed a paradigm that first extracts the patch features/tokens of the input image using a backbone, and then aggregates these patch features into a global descriptor via an aggregator. This backbone-plus-aggregator paradigm has achieved overwhelming dominance in the CNN era and remains widely used in transformer-based models. In this paper, however, we argue that a dedicated aggregator is not necessary in the transformer era, that is, we can obtain robust global descriptors only with the backbone. Specifically, we introduce some learnable aggregation tokens, which are prepended to the patch tokens before a particular transformer block. All these tokens will be jointly processed and interact globally via the intrinsic self-attention mechanism, implicitly aggregating useful information within the patch tokens to the aggregation tokens. Finally, we only take these aggregation tokens from the last output tokens and concatenate them as the global representation. Although implicit aggregation can provide robust global descriptors in an extremely simple manner, where and how to insert additional tokens, as well as the initialization of tokens, remains an open issue worthy of further exploration. To this end, we also propose the optimal token insertion strategy and token initialization method derived from empirical studies. Experimental results show that our method outperforms state-of-the-art methods on several VPR datasets with higher efficiency and ranks 1st on the MSLS challenge leaderboard. The code is available at https://github.com/lu-feng/image.