Fabrics Defect Dataset
An image dataset of fabric defects, that can be used for testing or training for AI models.
Asset Description
A set of fabric defects images that include, but are no limited to knots, thick/thin yarn, running shade variation, color yarn, holes, uneven dye, uneven print, crease marks, fabric skew, fabric browning, stain/spot/dirt, broken yarn, stop mark and more.
Dataset and Model Details
Dataset Overview
Images of fabric cutouts that contain defects that were found during a regular work day taken with a DSLR camera at a 12MPx resolution (4000 x 3000 px). The set includes top and bottom parts of the fabric and non-defective parts of the fabric. The resolution is around 10px/mm.
Usage
This is an image dataset and can be used for testing or training for AI models.
Maturity
Ongoing development: currently there are around 350 images where annotation has started based on findings from production. New images are added at a monthly basis.
Licence
Proprietary
Resources
- For further information, please contact eduard.modreanu@katty-fashion.ro with CC to project.management@katty-fashion.ro
Acknowledgement
The dataset was created in the framework of the EITM Teaching Factories project. In the AI REDGIO 5.0 project it will be used in the Industrial Pilot VII: Quality Assurance of clothing production at Katty Fashion The image dataset will be used in the future for training the AI system for the pilot to find fabric defects, not only product defects.