# RECOUNT: Reference-guided Counting with Synthetic Visual Exemplars

arXiv: https://arxiv.org/abs/2608.20621 · HTML: https://mlsanity.com/paper/2608.20621/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 27 of its 37 references that are in the map.
- First posted on arXiv 20 Aug 2026.

## Builds on (most cited first, 27 of 27)

- [Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection](https://mlsanity.com/paper/2303.05499/) (2023)
- [Vision Transformers Need Registers](https://mlsanity.com/paper/2309.16588/) (2023)
- [Eyes Wide Shut? Exploring the Visual Shortcomings of Multimodal LLMs](https://mlsanity.com/paper/2401.06209/) (2024)
- [When and why vision-language models behave like bags-of-words, and what to do about it?](https://mlsanity.com/paper/2210.01936/) (2022)
- [NWPU-Crowd: A Large-Scale Benchmark for Crowd Counting and Localization](https://mlsanity.com/paper/2001.03360/) (2020)
- [Learning To Count Everything](https://mlsanity.com/paper/2104.08391/) (2021)
- [Learning from Synthetic Data for Crowd Counting in the Wild](https://mlsanity.com/paper/1903.03303/) (2019)
- [JHU-CROWD++: Large-Scale Crowd Counting Dataset and A Benchmark Method](https://mlsanity.com/paper/2004.03597/) (2020)
- [CounTR: Transformer-based Generalised Visual Counting](https://mlsanity.com/paper/2208.13721/) (2022)
- [DINOv2 Meets Text: A Unified Framework for Image- and Pixel-Level Vision-Language Alignment](https://mlsanity.com/paper/2412.16334/) (2024)
- [A Low-Shot Object Counting Network With Iterative Prototype Adaptation](https://mlsanity.com/paper/2211.08217/) (2022)
- [CountGD: Multi-Modal Open-World Counting](https://mlsanity.com/paper/2407.04619/) (2024)
- [Zero-shot Object Counting](https://mlsanity.com/paper/2303.02001/) (2023)
- [Few-shot Object Counting and Detection](https://mlsanity.com/paper/2207.10988/) (2022)
- [DAVE -- A Detect-and-Verify Paradigm for Low-Shot Counting](https://mlsanity.com/paper/2404.16622/) (2024)
- [FLAIR: VLM with Fine-grained Language-informed Image Representations](https://mlsanity.com/paper/2412.03561/) (2024)
- [VLCounter: Text-aware Visual Representation for Zero-Shot Object Counting](https://mlsanity.com/paper/2312.16580/) (2023)
- [Point, Segment and Count: A Generalized Framework for Object Counting](https://mlsanity.com/paper/2311.12386/) (2023)
- [Training-free Object Counting with Prompts](https://mlsanity.com/paper/2307.00038/) (2023)
- [Zero-shot Object Counting with Good Exemplars](https://mlsanity.com/paper/2407.04948/) (2024)
- [From Fake to Real: Pretraining on Balanced Synthetic Images to Prevent Spurious Correlations in Image Recognition](https://mlsanity.com/paper/2308.04553/) (2023)
- [AFreeCA: Annotation-Free Counting for All](https://mlsanity.com/paper/2403.04943/) (2024)
- [Mind the Prompt: A Novel Benchmark for Prompt-based Class-Agnostic Counting](https://mlsanity.com/paper/2409.15953/) (2024)
- [CountGD++: Generalized Prompting for Open-World Counting](https://mlsanity.com/paper/2512.23351/) (2025)
- [Can Current AI Models Count What We Mean, Not What They See? A Benchmark and Systematic Evaluation](https://mlsanity.com/paper/2509.13939/) (2025)
- [From Open Set to Closed Set: Supervised Spatial Divide-and-Conquer for Object Counting](https://mlsanity.com/paper/2001.01886/) (2020)
- [FiGO: Fine-Grained Object Counting without Annotations](https://mlsanity.com/paper/2504.11705/) (2025)

## Abstract (arXiv)

Text-guided zero-shot object counters excel at spatial localization but categorize poorly on novel or fine-grained classes: natural language is too coarse to fully specify visual identity, so they fail to separate visually similar distractors. Few-shot counters sidestep this with visual exemplars, but require manual annotations on every image. To resolve this dilemma, we introduce RECOUNT, a plug-and-play framework for image-guided zero-shot counting. Rather than specify a category with a text prompt, our key insight is to specify it visually, from a single off-scene reference image. However, we find that a lone reference image provides narrow coverage of a category's appearance and is unreliable across diverse scenes. We therefore repurpose a diffusion model as an automated contrastive data engine that expands the reference into a diverse exemplar gallery, supplying the discriminative detail that text cannot. RECOUNT preserves the class-agnostic proposals of any frozen counter and offloads categorization to a separate visual module (a frozen backbone with a lightweight head trained on this synthetic data) that matches each proposal against the target and distractor galleries. Applied to a frozen counter, RECOUNT attains the best zero-shot accuracy on both benchmarks, cutting counting error (MAE) by 55% on LookAlikes and 21% on PairTally relative to the strongest prior zero-shot counter.