VQCounter: Designing Visual Prompt Queue for Accurate Open-World Counting

Ye, Fanfan; Fan, Yiqi; Zhong, Qiaoyong; Yang, Shicai; Xie, Di; Song, Jie; Song, Mingli

doi:10.24963/ijcai.2025/252

VQCounter: Designing Visual Prompt Queue for Accurate Open-World Counting

Fanfan Ye, Yiqi Fan, Qiaoyong Zhong, Shicai Yang, Di Xie, Jie Song, Mingli Song

Download BibTeX Cite

Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence. Main Track. Pages 2260-2268.

https://doi.org/10.24963/ijcai.2025/252

Keywords

Computer Vision: CV: Transfer, low-shot, semi- and un- supervised learning
Computer Vision: CV: Multimodal learning
Computer Vision: CV: Scene analysis and understanding

Abstract

Class-agnostic counting enables enumerating arbitrary object classes beyond those seen during training. Recent studies attempted to exploit the potential of visual foundation models such as GroundingDINO. Despite the considerable progress, we observe certain shortcomings, including the limited diversity of visual prompts and suboptimal training regimen.

To address these issues, we introduce VQCounter, which incorporates a visual prompt queue mechanism designed to enrich the diversity of visual prompts.

A random modality switching strategy is proposed during training to strengthen both textual and visual modalities.

Besides, in light of weak point supervision, a Voronoi diagram-based cost (VoronoiCost) is designed to improve Hungarian matching, leading to more stable and faster convergence.

Building upon the Voronoi diagram, we also propose a novel set of more stringent evaluation metrics, which take point localization into account.

Extensive experiments on the FSC-147 and CARPK datasets demonstrate that VQCounter achieves state-of-the-art performance in both zero-shot and few-shot settings, significantly outperforming existing methods across nearly all evaluations.

Fanfan Ye, Yiqi Fan, Qiaoyong Zhong, Shicai Yang, Di Xie, Jie Song, and Mingli Song. 2025. VQCounter: Designing Visual Prompt Queue for Accurate Open-World Counting. In Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence, IJCAI-25. IJCAI, 2260–2268. https://doi.org/10.24963/ijcai.2025/252

@inproceedings{ijcai2025p252,
  title     = {VQCounter: Designing Visual Prompt Queue for Accurate Open-World Counting},
  author    = {Ye, Fanfan and Fan, Yiqi and Zhong, Qiaoyong and Yang, Shicai and Xie, Di and Song, Jie and Song, Mingli},
  booktitle = {Proceedings of the Thirty-Fourth International Joint Conference on
               Artificial Intelligence, {IJCAI-25}},
  publisher = {International Joint Conferences on Artificial Intelligence Organization},
  editor    = {James Kwok},
  pages     = {2260--2268},
  year      = {2025},
  month     = {8},
  note      = {Main Track},
  doi       = {10.24963/ijcai.2025/252},
  url       = {https://doi.org/10.24963/ijcai.2025/252},
}

TY  - CPAPER
TI  - VQCounter: Designing Visual Prompt Queue for Accurate Open-World Counting
AU  - Ye, Fanfan
AU  - Fan, Yiqi
AU  - Zhong, Qiaoyong
AU  - Yang, Shicai
AU  - Xie, Di
AU  - Song, Jie
AU  - Song, Mingli
T2  - Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence, IJCAI-25
PB  - International Joint Conferences on Artificial Intelligence Organization
PY  - 2025
SP  - 2260
EP  - 2268
DO  - 10.24963/ijcai.2025/252
UR  - https://doi.org/10.24963/ijcai.2025/252
ER  -