README.md

---
language:
- en
license: cc-by-4.0
size_categories:
- 1K<n<10K
task_categories:
- video-text-to-text
tags:
- video-understanding
- temporal-reasoning
- counting
- benchmark
---

# VCBench: A Streaming Counting Benchmark for Spatial-Temporal State Maintenance in Long Videos

[**Paper**](https://huggingface.co/papers/2603.12703) | [**Code**](https://github.com/buaaplay/VCBench) | [**Dataset**](https://huggingface.co/datasets/buaaplay/VCBench)

VCBench is a streaming counting benchmark that repositions counting as a minimal probe for diagnosing **spatial-temporal state maintenance** capability in video-language models. By querying models at multiple timepoints during video playback, VCBench observes how model predictions evolve rather than checking isolated answers.

## Task Taxonomy

VCBench decomposes state maintenance into 8 fine-grained subcategories across two dimensions:

### Object Counting (tracking entities)
| Subcategory | Description |
|-------------|-------------|
| **O1-Snap** | How many objects are visible *at this moment*? |
| **O1-Delta** | How many objects appeared in the *past N seconds*? |
| **O2-Unique** | How many *different* individuals have appeared so far? |
| **O2-Gain** | How many *new* individuals appeared in the past N seconds? |

### Event Counting (tracking actions)
| Subcategory | Description |
|-------------|-------------|
| **E1-Action** | How many times has an atomic action occurred so far? |
| **E1-Transit** | How many scene transitions have occurred so far? |
| **E2-Episode** | How many activity segments have occurred so far? |
| **E2-Periodic** | How many complete cycles of a periodic action so far? |

## Dataset Summary

- **Total Videos**: 406 source videos (generating 4,574 clipped segments)
- **Total Size**: ~80 GB
- **Annotations**: 1,000 counting questions with 4,576 streaming query points and 10,071 frame-by-frame annotations.
- **Sources**: YouTube, ARKitScenes, ScanNet, ScanNet++, Ego4D, RoomTour3D, CODa, OmniWorld, and physics simulations.

## Usage

### Download via CLI

You can download the dataset using the `huggingface-cli`:

```bash
huggingface-cli download buaaplay/VCBench --repo-type dataset --local-dir data/videos
```

The `chunkedVideos/` directory contains 4,576 video clips (one per query point), each truncated to the query timestamp.

### Evaluation

To compute metrics (GPA, MoC, UDA) on results using the official evaluation scripts:

```bash
# Compute metrics on provided results
python eval/compute_metrics.py results/vcbench_gemini3flash_unified.jsonl data/vcbench_eval.jsonl
```

## Citation

```bibtex
@article{vcbench2025,
  title={VCBench: A Streaming Counting Benchmark for Spatial-Temporal State Maintenance in Long Videos},
  author={Liu, Pengyiang and Shi, Zhongyue and Hao, Hongye and Fu, Qi and Bi, Xueting and Zhang, Siwei and Hu, Xiaoyang and Wang, Zitian and Huang, Linjiang and Liu, Si},
  year={2026}
}
```

## License

This dataset and code are released under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/).