BDFD9: Bangladesh Centric Fabric Defect Dataset

BDFD9 is an image dataset of 1,088 knit and woven fabric photographs, captured on production lines at 4 garment factories in Bangladesh using a Google Pixel 6 and an Apple iPhone 14 Pro Max under ambient factory lighting. All images are original captures by the authors. Images are annotated in YOLOv8 format for 9 defect classes: DropStitch, Hole, Horizontal Mark, Knot, Oil Spot, Setup, Slub, Vertical LineMark, and Yarn Dust. Data is split into train (775), valid (208), and test (105) images, and organized by defect class within each split (images/ and labels/ per class). Images with multiple defect types are duplicated across each relevant class folder. dataset_metadata.csv lists each image's split, classes, and box count.