Back to datasets
Dataset assetOpen Source CommunityMachine LearningComputer Vision
Mahjong Dataset
A computer‑vision dataset for Chinese Mahjong tiles, containing various tile images and their labels for training and testing machine‑learning models.
Source
github
Created
Apr 29, 2019
Updated
May 12, 2024
Signals
554 views
Availability
Linked source ready
Overview
Dataset description and usage context
Mahjong Dataset Overview
Dataset Structure
./raw-images: Folder with original, un‑cropped images../raw-tiles: Folder with uncropped tile images (no scaling)../tiles-resized: Folder with scaled, ready‑to‑use tile images../tiles-data: CSV file containing image labels and its template folder../untagged-images-raw: Folder with original, un‑cropped images without tags../untagged-tiles: Folder with untagged tile images.
Data Description
- Image Sources: Mainly collected from Google Image Search, eBay, and Alibaba.
- Scaled Image Size: 240 (W) × 320 (H),
.jpgformat.
Dataset Content
train.zip: Pre‑packaged dataset at the repository root, containing:imagesfolder: Mahjong tile images.data.csv: Image label data.
Label Format
| image-name | label | label-name |
|---|---|---|
| 1.jpg | 38 | bonus‑winter |
| ... | ... | ... |
Category Information
| table-name | table-index |
|---|---|
| dots-1 | 1 |
| dots-2 | 2 |
| dots-3 | 3 |
| dots-4 | 4 |
| dots-5 | 5 |
| dots-6 | 6 |
| dots-7 | 7 |
| dots-8 | 8 |
| dots-9 | 9 |
| bamboo-1 | 10 |
| bamboo-2 | 11 |
| bamboo-3 | 12 |
| bamboo-4 | 13 |
| bamboo-5 | 14 |
| bamboo-6 | 15 |
| bamboo-7 | 16 |
| bamboo-8 | 17 |
| bamboo-9 | 18 |
| characters-1 | 19 |
| characters-2 | 20 |
| characters-3 | 21 |
| characters-4 | 22 |
| characters-5 | 23 |
| characters-6 | 24 |
| characters-7 | 25 |
| characters-8 | 26 |
| characters-9 | 27 |
| honors-east | 28 |
| honors-south | 29 |
| honors-west | 30 |
| honors-north | 31 |
| honors-red | 32 |
| honors-green | 33 |
| honors-white | 34 |
| bonus-spring | 35 |
| bonus-summer | 36 |
| bonus-autumn | 37 |
| bonus-winter | 38 |
| bonus-plum | 39 |
| bonus-orchid | 40 |
| bonus‑chrysanthemum | 41 |
| bonus‑bamboo | 42 |
License
The dataset is released under the MIT License.
Need downstream help?
Pair the dataset with AI analysis and content workflows.
Once the source passes your review, move straight into summarization, transformation, report drafting, or presentation generation with the JuheAI toolchain.