2026-10-15T09:47:09Z https://data.mendeley.com/oai

oai:data.mendeley.com/8y46pm2m6y.1 2026-06-18T10:18:51Z

Islam, Md Riadul TILDA-OBB Mendeley Data The annotated dataset, TILDA-OBB, contains 1,562 images across four fabric groups: Plain (C1), Texture (C2), Printed (C3), and Patterned (C4). It is derived from the TILDA Textile Texture Database (Schulz-Mirbach, 1996), developed at Technische Universität Hamburg-Harburg under the DFG research programme "Automatic Visual Inspection of Technical Objects." TILDA contains 3,200 grayscale images (768×512) across eight fabric types and seven defect classes, plus one defect-free class (Albert-Ludwigs-Universität Freiburg, 1995). The raw images carry only short German text labels with no bounding-box annotations; these were translated into English and annotated in Roboflow (Dwyer and Nelson, 2024) using oriented bounding boxes. Five defect classes were retained: Cut, Foreign Object, Hole, Stain, and Thread Defect. The No Defect and Wrinkle classes were excluded, as neither contains a spatially distinguishable defect region. Textile Engineering Cotton Fabrics Dataset doi:10.17632/8y46pm2m6y.1 oai:data.mendeley.com/8y46pm2m6y.1 Creative Commons Attribution 4.0 International http://creativecommons.org/licenses/by/4.0 https://data.mendeley.com/datasets/8y46pm2m6y 2026-06-18T10:18:51Z