---
license: apache-2.0
annotations_creators:
- expert-generated
- machine-generated
language:
- en
size_categories:
- 100K<n<1M
task_categories:
- video-classification
pretty_name: GRW — Gesture Recognition in-the-Wild
tags:
- co-speech gestures
- semantic gestures
- gesture-recognition
- temporal-localization
- video-understanding
- multimodal-learning
configs:
- config_name: semantic_classification
  data_files:
  - split: train
    path: semantic_classification_train.csv
  - split: test
    path: semantic_classification_test.csv
  - split: test_unseen_words
    path: semantic_classification_test_unseen_words.csv
- config_name: recognition_localization
  data_files:
  - split: train
    path: recognition_localization_train.csv
  - split: test
    path: recognition_localization_test.csv
---

# GRW: Gesture Recognition in-the-Wild

This dataset is associated with the paper **"Recognizing Co-speech Gestures in-the-Wild"** (ECCV 2026).

<p align="center">
  <a href="https://arxiv.org/abs/2605.31589"><img src="https://img.shields.io/badge/arXiv-2605.31589-b31b1b.svg" alt="arXiv"></a>
  <a href="https://www.robots.ox.ac.uk/~vgg/research/grw/"><img src="https://img.shields.io/badge/Project-Page-blue.svg" alt="Project Page"></a>
  <a href="https://www.robots.ox.ac.uk/~vgg/research/grw/dataset"><img src="https://img.shields.io/badge/Explore-Dataset-green.svg" alt="Explore the dataset"></a>
  <a href="https://github.com/Sindhu-Hegde/grw"><img src="https://img.shields.io/badge/Code-GitHub-black.svg" alt="Code"></a>
  <a href="https://opensource.org/licenses/Apache-2.0"><img src="https://img.shields.io/badge/License-Apache_2.0-yellow.svg" alt="License: Apache 2.0"></a>
</p>

<p align="center">
  <img src="https://raw.githubusercontent.com/Sindhu-Hegde/grw/main/assets/teaser.gif" width="600" alt="Semantic co-speech gestures recognized and localized in real-world videos"/>
</p>

Our aim is to recognise and localize **semantic gestures** in real-world videos. These gestures are visually depictive and semantically linked to specific spoken words. We introduce a new large-scale benchmark, **GRW** (Gesture Recognition in-the-Wild), which provides word-level annotations and gesture boundaries for semantic gestures occurring in unconstrained real-world settings.

---

## 📋 Table of Contents

- [📚 What is GRW?](#📚-what-is-grw)
- [⚡ Quickstart](#⚡-quickstart)
- [📦 Getting the data](#📦-getting-the-data)
- [📝 Dataset structure](#📝-dataset-structure)
  - [Task 1: Semantic classification](#task-1-semantic-classification)
  - [Task 2: Gesture recognition & localization](#task-2-gesture-recognition--localization)
  - [Data instances](#data-instances)
- [📊 Statistics](#📊-statistics)
- [🏆 Benchmark results](#🏆-benchmark-results)
- [🔖 Citation](#🔖-citation)
- [📧 Contact](#📧-contact)
- [⚖️ License & data usage](#⚖️-license--data-usage)

---

## 📚 What is GRW?

GRW is a large-scale video dataset for gesture recognition in the wild, containing over **17,000 semantic gesture clips** spanning **155 gesture words**. Each video clip is **4 seconds** long and features a single speaker performing a co-speech gesture. The dataset covers a wide range of gesture categories, including iconic, deictic, metaphoric, and beat gestures, across diverse speakers, backgrounds, and real-world settings. In addition to word-level speech boundaries, the dataset provides **manually annotated temporal gesture boundaries** for each video clip.

<div align="center">

| **17K+** | **155** | **3** |
|:--:|:--:|:--:|
| Semantic gesture clips | Word vocabulary | Modalities<br>Video · Speech · Text |

</div>

The dataset supports two tasks:
| # | Task | Given | Predict |
|:--|:--|:--|:--|
| **1** | **Semantic classification** | A candidate video segment and a spoken target word | Whether the segment contains a gesture semantically related to that word |
| **2** | **Gesture recognition & localization** | A gesture video clip | The gestured word, and the temporal gesture boundaries |

> 🎬 **Browse the videos** on the [interactive dataset explorer](https://www.robots.ox.ac.uk/~vgg/research/grw/dataset) before downloading.

---

## ⚡ Quickstart

This repository contains the **annotation CSV files**. The videos themselves are sourced from YouTube and obtained separately — see [Getting the data](#📦-getting-the-data).

```python
from datasets import load_dataset

# Task 1: Semantic classification
sem = load_dataset("sindhuhegde/grw", "semantic_classification")
# splits: "train", "test", "test_unseen_words"

# Task 2: Gesture recognition & localization
rec = load_dataset("sindhuhegde/grw", "recognition_localization")
# splits: "train", "test"

print(rec["test"][0])
```

Or read the CSVs directly with `pandas`:

```python
import pandas as pd
from huggingface_hub import hf_hub_download

path = hf_hub_download("sindhuhegde/grw",
                       "recognition_localization_test.csv",
                       repo_type="dataset")
df = pd.read_csv(path)
```

<details>
<summary><b>Files in this repository</b></summary>

| File | Task | Split | Rows |
|:--|:--|:--|--:|
| `semantic_classification_train.csv` | Semantic classification | train | 135,503 |
| `semantic_classification_test.csv` | Semantic classification | test | 4,000 |
| `semantic_classification_test_unseen_words.csv` | Semantic classification | test (unseen words) | 500 |
| `recognition_localization_train.csv` | Recognition & localization | train | 15,340 |
| `recognition_localization_test.csv` | Recognition & localization | test | 2,000 |

</details>

---

## 📦 Getting the data

The CSVs give you YouTube IDs, timestamps and speaker bounding boxes. There are **two ways** to obtain the data — pick whichever suits you.

### Option A — Download pre-extracted SHuBERT features *(recommended)*

Our gesture models take SHuBERT features as input, and we release them directly. **This skips video downloading and preprocessing entirely.**

| Feature set | Download |
|:--|:--|
| Semantic Classification — Train | [Link](https://thor.robots.ox.ac.uk/grw/shubert-features/semantic_classification_train.tar) |
| Semantic Classification — Test | [Link](https://thor.robots.ox.ac.uk/grw/shubert-features/semantic_classification_test.tar) |
| Semantic Classification — Test unseen words | [Link](https://thor.robots.ox.ac.uk/grw/shubert-features/semantic_classification_test_unseen_words.tar) |
| Word Recognition & Localization — Train | [Link](https://thor.robots.ox.ac.uk/grw/shubert-features/recognition_localization_train.tar) |
| Word Recognition & Localization — Test | [Link](https://thor.robots.ox.ac.uk/grw/shubert-features/recognition_localization_test.tar) |

Checksums are available [here](https://thor.robots.ox.ac.uk/grw/shubert-features/SHA512SUMS) and can be verified with `sha512sum -c SHA512SUMS`.

<details>
<summary><b>Folder structure of the extracted features</b></summary>

```text
shubert_features (path of the extracted shubert features)
├── semantic_classification
│   ├── split (<train>/<test>/<test_unseen_words>)
│   │   ├── *.npy
├── recognition_localization
│   ├── split (<train>/<test>)
│   │   ├── *.npy
```

</details>

### Option B — Download and crop the videos yourself

Use the preprocessing scripts in the [code repository](https://github.com/Sindhu-Hegde/grw):

```bash
git clone https://github.com/Sindhu-Hegde/grw.git
cd grw/preprocess

# Download the videos from YouTube-ids and timestamps
python download_videos.py --input_csv=<csv-file> --result_dir=<raw-video-root>

# Crop the videos with the bounding-box co-ordinates provided in the csv files
python crop_videos.py --input_csv=<csv-file> --video_dir=<raw-video-root> --output_dir=<preprocessed-video-root>
```

> ⚠️ **Note:** Due to new YouTube policies, downloading videos (especially for the train sets) might take a long time. If you only need to train or evaluate the gesture models, prefer **Option A**.

<details>
<summary><b>Folder structure after download and pre-processing</b></summary>

```text
raw_video_root (path of the downloaded raw videos)
├── *.mp4 (raw uncropped videos)
```

```text
preprocessed_video_root (path of the pre-processed gesture videos)
├── word folders
│   ├── *.mp4 (extracted person-specific gesture video)
```

</details>

---

## 📝 Dataset structure

All frame indices are given **at 25 fps**.

### Task 1: Semantic classification

Video segment annotations for classifying whether a candidate segment contains a gesture that is semantically related to the spoken target word.

`semantic_classification_train.csv` · `semantic_classification_test.csv` · `semantic_classification_test_unseen_words.csv`

| Column | Description |
|:--|:--|
| `fname` | unique video clip identifier |
| `target_word` | gesture word label |
| `speech_start` / `speech_end` | speech boundaries (in frames at 25fps), automatically extracted using WhisperX |
| `context_start` / `context_end` | frame boundaries (at 25fps) of the context window preceding the candidate segment |
| `target_start` / `target_end` | frame boundaries (at 25fps) of the candidate segment being classified |
| `word_form` | surface form of the word as spoken |
| `gesture_label` | `1` if the candidate segment contains a gesture semantically related to the target word, `0` otherwise |
| `source_file` | YouTube video ID |
| `source_video_start` / `source_video_end` | start and end timestamps (in seconds) of the clip within the source video |
| `source_width` / `source_height` | resolution (in pixels) of the source video |
| `num_frames` | total number of frames in the original (long) source video; provided for preprocessing |
| `pad` / `bbox` | padding (in pixels) and bounding box `[x1, y1, x2, y2]` (in pixels) computed on the original source video; used internally to crop the speaker region during preprocessing |

### Task 2: Gesture recognition & localization

Video segment annotations with manually annotated gesture boundaries, for gesture recognition and temporal localization.

`recognition_localization_train.csv` · `recognition_localization_test.csv`

| Column | Description |
|:--|:--|
| `fname` | unique video clip identifier |
| `target_word` | gesture word label |
| `speech_start` / `speech_end` | speech boundaries (in frames at 25fps), automatically extracted using WhisperX |
| `gesture_start` / `gesture_end` | **manually annotated** gesture boundaries (in frames at 25fps) |
| `word_form` | surface form of the word as spoken |
| `source_file` | YouTube video ID |
| `source_video_start` / `source_video_end` | start and end timestamps (in seconds) of the clip within the source video |
| `source_width` / `source_height` | resolution (in pixels) of the source video |
| `num_frames` | total number of frames in the original (long) source video; provided for preprocessing |
| `pad` / `bbox` | padding (in pixels) and bounding box `[x1, y1, x2, y2]` (in pixels) computed on the original source video; used internally to crop the speaker region during preprocessing |

> 🔒 For **both** tasks, the train and test sets have **disjoint videos**.

### Data instances

<details open>
<summary><b>Recognition & localization</b></summary>

```json
{
  "fname": "entire/00000",
  "target_word": "entire",
  "speech_start": 35,
  "speech_end": 44,
  "gesture_start": 25,
  "gesture_end": 67,
  "word_form": "entire",
  "source_file": "7IwTua4AaRY",
  "source_video_start": 58.36,
  "source_video_end": 62.36,
  "source_width": 1920,
  "source_height": 1080,
  "num_frames": 282,
  "pad": 454,
  "bbox": "[1259, 521, 2168, 1430]"
}
```

</details>

<details>
<summary><b>Semantic classification (a <code>gesture_label = 0</code> example)</b></summary>

```json
{
  "fname": "look/00012",
  "target_word": "look",
  "speech_start": 197,
  "speech_end": 201,
  "context_start": 0,
  "context_end": 149,
  "target_start": 150,
  "target_end": 249,
  "word_form": "look",
  "gesture_label": 0.0,
  "source_file": "ogCJrrvgais",
  "source_video_start": 1054.48,
  "source_video_end": 1058.48,
  "source_width": 1280,
  "source_height": 720,
  "num_frames": 1725,
  "pad": 426,
  "bbox": "[727, 384, 1579, 1236]"
}
```

</details>

Use the [dataset viewer](https://huggingface.co/datasets/sindhuhegde/grw/viewer) to explore more examples.

---

## 📊 Statistics

Every clip is 4 seconds long. All counts below are computed directly from the released CSV files.

### Task 1: Semantic classification

| Split | # Rows | # Words | # Source videos | Gestured (`1`) | Not gestured (`0`) |
|:--|--:|--:|--:|--:|--:|
| `train` | 135,503 | 155 | 34,132 | 15,340 | 120,163 |
| `test` | 4,000 | 100 | 3,297 | 2,000 | 2,000 |
| `test_unseen_words` | 500 | 10 | 468 | 201 | 299 |

The `test_unseen_words` split evaluates generalization to words **never seen during training** — its 10 words have zero overlap with the 155 training words: `crawl`, `direct`, `enlarge`, `great`, `ingest`, `loads`, `proximity`, `rearrange`, `synchronize`, `uniform`.

### Task 2: Gesture recognition & localization

| Split | # Clips | # Words | # Source videos | Clips per word (median) |
|:--|--:|--:|--:|--:|
| `train` | 15,340 | 155 | 10,407 | 65 |
| `test` | 2,000 | 100 | 1,435 | 14 |

All 100 test words are contained within the 155 training words. In the training set, the annotated gestures last **40 frames on average (≈1.6 s)** out of the 4-second clip.

### Vocabulary

<details>
<summary><b>The 155 gesture words</b></summary>

```
above, absorb, angle, arc, around, ascend, back, balance, barrier, beautiful, begin,
below, big, block, boost, bottom, bounce, branch, break, broad, build, bundle, bye,
call, catch, circle, close, collect, collide, combine, compress, condense, connect,
count, cross, cup, curve, decrease, deep, descend, develop, direction, down, eat,
elevate, embrace, engage, entire, evolve, expand, explode, few, fight, five, flip,
flow, focus, force, four, front, full, gigantic, global, grab, grasp, grow, hashtag,
heavy, height, hello, her, high, hold, horizontal, hug, huge, increase, interaction,
join, knock, large, layer, less, lift, link, little, long, look, loop, lower, many,
merge, mix, move, narrow, no, open, overlap, pause, peak, perfect, pieces, point,
press, process, push, quick, raise, reduce, roll, rotate, round, run, separate,
shake, she, short, shrink, slide, small, specific, spin, spiral, stack, stop,
straight, stretch, strong, switch, three, throw, tie, tight, tilt, tiny, together,
top, track, transform, transition, trap, turn, twist, two, unify, us, various, wait,
walk, wave, whole, wide, wrap, yes, zoom
```

The full list is also browsable on the [Word List page](https://www.robots.ox.ac.uk/~vgg/research/grw/dataset/#/wordlist).

</details>

`target_word` is the canonical class label, while `word_form` records how the word was actually spoken — so inflections are preserved (e.g. `quick` → `quickly`, `wave` → `waving`, `rotate` → `rotating`). The training set contains 381 distinct `target_word` / `word_form` pairs.

---

## 🏆 Benchmark results

Results of our models on the GRW test sets, reproducible with the [evaluation scripts](https://github.com/Sindhu-Hegde/grw#-evaluation).

<div align="center">

*Semantic gesture classification on the GRW test set*
| Accuracy | Precision | Recall | High-confidence Accuracy |
|:--:|:--:|:--:|:--:|
| **75.83** | **79.91** | **69.00** | **93.20** |

*Word recognition and localization on the GRW test set*
| Acc@1 | Acc@5 | Acc@10 | mIoU |
|:--:|:--:|:--:|:--:|
| **18.35** | **37.30** | **51.70** | **0.67** |

</div>

---

## 🔖 Citation

If you find this dataset helpful, please consider starring ⭐ the [repository](https://github.com/Sindhu-Hegde/grw) and citing our work.

```bibtex
@inproceedings{hegde_eccv_2026,
  title={Recognizing Co-Speech Gestures in-the-Wild},
  author={Hegde, Sindhu and Prajwal, K R and Zisserman, Andrew},
  booktitle={European Conference on Computer Vision (ECCV)},
  year={2026}
}
```

---

## 📧 Contact

For questions about the dataset, access requests, or collaboration inquiries, please email **sindhu@robots.ox.ac.uk**.

| Author | Affiliation |
|:--|:--|
| [Sindhu Hegde](https://sindhu-hegde.github.io/) | University of Oxford |
| [K R Prajwal](https://www.robots.ox.ac.uk/~prajwal/) | University of Oxford |
| [Andrew Zisserman](https://scholar.google.com/citations?hl=en&user=UZ5wscMAAAAJ) | University of Oxford |

Visual Geometry Group (VGG) · Department of Engineering Science · University of Oxford

---

## ⚖️ License & data usage

The annotations in this repository are released under the **Apache 2.0** licence. The underlying videos are sourced from YouTube and are not redistributed here — they remain subject to their original terms of use. The dataset is intended for research purposes.