LangFabric: 6,384 Language-Annotated Fabric Images for Defect Detection

6,384 paired samples of high-resolution fabric images and structured textual descriptions, covering 16 fabric textures and 12 defect types. The dataset, created by Yan-Qin Ni and last updated in April 2026, includes pixel-level segmentation masks and was collected inline from an industrial air-jet loom under realistic manufacturing conditions. LangFabric is designed as a benchmark for multimodal fabric defect analysis.

Use Cases

• Supervised fabric defect detection based on high-resolution images and segmentation masks
• Anomaly detection and one-class learning based on 12 distinct defect types
• Vision-language model alignment research based on structured textual prompts describing defect attributes
• Industrial inspection system development based on data from realistic manufacturing conditions

Strengths

• Contains 6,384 paired image-text samples, providing a substantial multimodal resource
• Covers 16 distinct fabric texture categories and 12 defect types, including structural and non-structural anomalies
• Includes pixel-level semantic segmentation masks for precise defect localization
• Data was collected inline from an industrial air-jet loom, ensuring practical variability

Limitations

• Column-level documentation is absent; field semantics must be inferred after download
• Row count is unknown, which may limit suitability assessment

Provenance

Source

Yan-Qin Ni via figshare

Collection Method

Collected inline from an industrial air-jet loom under realistic manufacturing conditions

Time Range

null

Freshness

Last updated 2026-04 09:21:16

Geography

null