---
license: cc-by-4.0
task_categories:
- image-segmentation
tags:
- fabric
- textile
pretty_name: FabWrinkle-13
size_categories:
- 10K<n<100K
viewer: false
---

# FabWrinkle-13

### Description

FabWrinkle-13, a newly generated dataset of 15,187 synthetic images covering 13 different types of plain and patterned fabrics. The dataset was created using a synthetic data generation pipeline that employs OpenGL with Gouraud shading, Lambertian illumination, and random camera positions. Wrinkles were generated by dropping a cloth mesh from a height onto random obstacles.

Folders are named by the category of the fabric. In each folder there are 3 files for each data. 
* The fabric image - type: png, filename: "unique id", ex: 10522300.png
* JSON file with meta data, corner information, used obstacles - type: json, filename: "same unique id", ex: 10522300.json
* The wrinkle mask - type:png, filename: "same unique id"-d, ex: 10522300-d.png

### Fabric Image Counts

| Fabric Name | Image Count |
|-------------|-------------|
| bubble      |   1053      |
| camo        |   1255      |
| floral_1    |   1232      |
| floral_2    |   1204      |
| floral_3    |   934       |
| msu_1       |   1009      |
| msu_2       |   1012      |
| plain_1     |   2101      |
| plain_2     |   1030      |
| plain_3     |   1067      |
| plain_4     |   1048      |
| plain_5     |   1124      |
| vehicle     |   1120      |

### Citation