RoboCloth: Massive Real-Cloth Dataset With 80 Billion Observations Released

Researchers from SFU, Westlake University and others released RoboCloth, a large-scale real-cloth material dataset with 500 fabric types and 80 billion observations for neural reflectance reconstruction, ending the field's reliance on the 2003 MERL BRDF data. First author is SFU GrUVi Lab master's student Zhen (Colin) Li.

2026-10-20 ~ 2026-10-20 · 4 related posts

• RoboCloth: 500 real cloth materials, 80B observations for neural reflectance reconstruction — taiyasaki · 2026-10-20
• RoboCloth: 500 Fabrics, 80B Observations Break 2003 MERL BRDF Data Reliance — taiyasaki · 2026-10-20
• RoboCloth: 500 real cloths, 80B observations to replace 2003-era BRDF data — taiyasaki · 2026-10-20
• RoboCloth advisor: the dataset was 'much harder than it looks' — taiyasaki · 2026-10-20