Data Bite
KolektorSDD2
A surface-quality inspection dataset for studying defect segmentation under sparse and imbalanced manufacturing defect conditions.
Manufacturing
Surface defect segmentation
CC BY-SA 4.0
KolektorSDD2 focuses on surface-quality inspection in manufacturing. It is useful for studying settings where defective samples are much rarer than normal samples and where the target is not only classification, but also localizing the defective region.
The dataset is a good fit for experiments around class imbalance, small defects, and segmentation masks in industrial inspection workflows.
Dataset facts
- Use for: surface defect segmentation, manufacturing inspection, imbalanced defect detection experiments
- Scale: Dataset Ninja lists 3,336 images and 15,764 annotated defect objects
- Watch out: Dataset Ninja lists CC BY-SA 4.0; check the original citation and dataset terms before redistribution.