This dataset is an official version 2 of the MISTT Dataset originally introduced in the paper:
Fine-grained Stroke Recognition in Broadcast Table Tennis Videos with ATDT (ACM TOMM 2025).
This repository provides a version 2 video database built upon the foundations of the MISTT project. It incorporates a total of 24 world-class table tennis matches (including WTT Smashes, ITTF World Championships, and the Tokyo Olympics).
Below is the detailed taxonomy of table tennis strokes annotated in this dataset, categorized by Forehand (正手) and Backhand (反手) actions along with their corresponding English technical terms:
| English Technical Term | Forehand (正手) | Backhand (反手) |
|---|---|---|
| Drop shot / Long push | 擺短 / 劈長 | 擺短 / 劈長 |
| Sidespin counter | 晃 | 撇 |
| Flick / Chiquita | 撥球 (挑球) | 擰拉 |
| Spin / Counterdrive / Counterloop | 拉上旋 / 反拉 / 對拉 | 拉上旋 / 反拉 / 對拉 |
| Loop | 拉下旋 | 拉下旋 |
| Lob | 放高球 | 放高球 |
| Attack / Block / Fast drive (loop) / Slam | 攻 / 擋 / 快帶 / 殺球 | 攻 / 擋 / 反撕 / 彈 (殺球) |
| Serve | 發球 | 發球 |
| Other | 其他 | 其他 |
ACTION_NAMES_28C = {
# Forehand (正手)
0: "正-擺短", # Forehand - Drop shot
1: "正-劈長", # Forehand - Long push
2: "正-晃/撇", # Forehand - Sidespin counter
3: "正-撥球/擰拉", # Forehand - Flick
4: "正-拉(下)", # Forehand - Loop
5: "正-拉(上)", # Forehand - Spin
6: "正-攻", # Forehand - Attack
7: "正-擋", # Forehand - Block
8: "正-快帶/反撕", # Forehand - Fast drive
9: "正-反拉", # Forehand - Counterdrive
10: "正-對拉", # Forehand - Counterloop
11: "正-放高球", # Forehand - Lob
12: "正-殺球", # Forehand - Slam
13: "正-發球", # Forehand - Serves
# Backhand (反手)
14: "反-擺短", # Backhand - Drop shot
15: "反-劈長", # Backhand - Long push
16: "反-晃/撇", # Backhand - Sidespin counter
17: "反-撥球/擰拉", # Backhand - Chiquita
18: "反-拉(下)", # Backhand - Loop
19: "反-拉(上)", # Backhand - Spin
20: "反-攻", # Backhand - Attack
21: "反-擋", # Backhand - Block
22: "反-快帶/反撕", # Backhand - Fast loop
23: "反-反拉", # Backhand - Counterdrive
24: "反-對拉", # Backhand - Counterloop
25: "反-放高球", # Backhand - Lob
26: "反-殺球", # Backhand - Slam
27: "反-發球", # Backhand - Serves
}To facilitate direct model training and inference on downloaded YouTube videos, we provide both raw and adjusted frame-aligned annotations:
- Raw Annotations (
labeled_json/): Contains the original human-annotated stroke timestamps based on the native broadcast acquisition frame rates. - YouTube-Aligned Annotations (
YT_json/): To resolve frame rate discrepancies caused by video downloaders, all annotations in this directory have been converted and frame-aligned to match standard YouTube playback frame rates.
Note: Due to conversion and rounding, there may be a minor boundary offset of a few frames compared to the raw video.
Note: We provide scripts to label player positions (left vs. right), which are available for use if required.
To help researchers quickly understand our data structure, below is a real snippet from our annotation label files (e.g., {match}_data.txt).
Player EventType 正/反拍 動作類型 StartFrame Duration Index
---------------------------------------------------------------------------
L 發球 正 發球 39 59 0
R 動作 反 撥球(挑球)/擰拉 88 22 2
L 動作 正 拉(上旋) 98 22 4
R 動作 反 反拉 109 20 6
L 動作 正 拉(上旋) 119 22 8
R 動作 反 攻 128 24 10
L 動作 反 反拉 141 23 12
R 動作 反 反拉 151 22 14
L 動作 反 反拉 164 19 16
| Index | Match | Video ID | YouTube Link | Annotation FPS | Convert FPS |
|---|---|---|---|---|---|
| 1 | 2022WTT Smash-成年男子組, 馬龍 VS 林昀儒 | CmbwzmO6PX | YouTube | 25 fps | 50 fps |
| 2 | 2022WTT Star Contender Doha-成年男子組, 林昀儒 VS 林鐘勳 | odugj313m2 | YouTube | 25 fps | 25 fps |
| 3 | 2022WTT Star Contender ESS-成年男子組, 林昀儒 VS 王楚欽 | 65a7fSWxre | YouTube | 29.97 fps | 29.97 fps |
| 4 | 2022WTT冠軍賽布達佩斯站-成年男子組, Hugo Calderano VS 林鐘勳 | i6n6xjDxCk | YouTube | 25 fps | 50 fps |
| 5 | 2022WTT冠軍賽布達佩斯站-成年男子組, 莊智淵 VS 梁靖崑 | DbIkSN2KxR | YouTube | 25 fps | 50 fps |
| 6 | 2021世界桌球錦標賽-成年男子組, 樊振東 VS 林高遠 | Da3eHy0HU2 | YouTube | 30 fps | 50 fps |
| 7 | 2022WTT冠軍賽布達佩斯站-成年男子組, 馬龍 VS Patrick Franziska | dhqtw1pGe1 | YouTube | 25 fps | 50 fps |
| 8 | 2021世界桌球錦標賽-成年男子組, 梁靖崑 VS Hugo Calderano | Tj7hSZvhqW | YouTube | 30 fps | 50 fps |
| 9 | 2022WTT冠軍賽布達佩斯站-成年女子組, 孫穎莎 VS 付玉 | Bddr8bz4eH | YouTube | 25 fps | 50 fps |
| 10 | 2021世界桌球錦標賽-成年男子組, 樊振東 VS Moregardh | 7O6uCFgpq8 | YouTube | 25 fps | 50 fps |
| 11 | 2022WTT冠軍賽布達佩斯站-成年女子組, 陳思羽 VS 伊藤美誠 | wMhfRhWN6S | YouTube | 25 fps | 50 fps |
| 12 | 2022WTT冠軍賽布達佩斯站-成年男子組, Timo Boll VS 林高遠 | zzuX0uFxIj | YouTube | 25 fps | 50 fps |
| 13 | 2022WTT冠軍賽布達佩斯站-成年女子組, 陳夢 VS 王曼昱 | 4yYvMIVQNZ | YouTube | 30 fps | 30 fps |
| 14 | 2022WTT冠軍賽布達佩斯站-成年男子組, Patrick Franziska VS 張本智和 | dgVLPQOFbr | YouTube | 25 fps | 50 fps |
| 15 | 2022WTT冠軍賽布達佩斯站-成年男子組, 莊智淵 VS 林高遠 | AuhW1I2dZQ | YouTube | 25 fps | 50 fps |
| 16 | 2022世界桌球團體錦標賽-成年女子組, 王曼昱 VS 伊藤美誠 | nYoaxjNLXd | YouTube | 30 fps | 30 fps |
| 17 | 2022WTT冠軍賽布達佩斯站-成年男子組, 張本智和 VS 林鐘勳 | uPQ2hRtPim | YouTube | 25 fps | 50 fps |
| 18 | 2021WTT Cup Final Singapore-成年男子組, 林昀儒 VS 王楚欽 | cSM5vGjmEm | YouTube | 25 fps | 50 fps |
| 19 | 2021WTT Cup Final Singapore-成年男子組, 樊振東 VS 王楚欽 | 4tYoJuM6If | YouTube | 25 fps | 50 fps |
| 20 | 2021WTT Cup Final Singapore-成年男子組, 樊振東 VS 張本智和 | Dlh9nMDYzm | YouTube | 25 fps | 50 fps |
| 21 | 2022世界桌球團體錦標賽-成年男子組, 王楚欽 VS 張本智和 | eXZe1cgNmk | YouTube | 30 fps | 30 fps |
| 22 | 2022Singapore Smash-成年男子組, 馬龍 VS 王楚欽 | Lu9I2SrfHe | YouTube | 25 fps | 50 fps |
| 23 | 2022WTT Cup Finals XinXiang-成年女子組, 王曼昱 VS 陳夢 | aNgTF2Dol5 | YouTube | 30 fps | 30 fps |
| 24 | 2022WTT冠軍賽布達佩斯站-成年男子組, 張本智和 VS 林高遠 | ijzGS7HBBS | YouTube | 25 fps | 50 fps |
If you find this dataset useful in your research, please consider citing the original paper and this v2 dataset 😎:
@article{chang2025fine,
title={Fine-grained Stroke Recognition in Broadcast Table Tennis Videos with ATDT},
author={Chang, Tang-Chen and Jheng, Duen-Chian and Liang, Hsuan-Ya and Harchan, Bill Louis and Ching, Pu and Tsai, Tsung-Hsun and Chang, Chih-Yi and Wu, Te-Cheng and Li, Yung-Hui and Pan, Tse-Yu and others},
journal={ACM Transactions on Multimedia Computing, Communications and Applications},
volume={21},
number={12},
pages={1--24},
year={2025},
publisher={ACM New York, NY}
}