//
Category
Enterprise AI
//
Stack
Industrial Intelligence
//
Timeline
5 Months
// problem
A leading apparel exporter in Sri Lanka faced rising fabric defect rates that threatened export quality standards. Processing hundreds of fabric rolls daily for global brands, the organization relied on manual fabric inspection methods that struggled to keep up with both production volume and defect diversity. Issues like slubs, holes, oil stains, and weft distortions were frequently missed by human inspectors—especially during continuous shifts. This led to shipment rejections, customer complaints, and production delays, resulting in annual losses worth millions.
// How we fixed it
We built Fabrin, a real-time AI-powered fabric inspection platform capable of detecting defects as small as 0.1mm. Fabrin integrates directly into the fabric rolling machines and inspects fabric at production speed using high-resolution cameras and precision lighting. The system identifies defects, categorizes them, and instantly alerts operators. It also auto-generates batch-wise reports for compliance and quality assurance.
// How we made it happen
Our team thoroughly assessed the client’s textile production line, examining fabric types, lighting, and inspection challenges. We developed a plug-and-play solution integrating industrial-grade line-scan cameras, AI-powered edge computing units, and PLC connectivity with the fabric winder. The dual-stage AI model combines unsupervised anomaly detection and supervised defect classification, continuously improving via active learning. A centralized dashboard allows QA managers to monitor defects, adjust tolerances, and export reports to ERP systems. The solution was successfully deployed across four production units within two months, trained to identify over 30 fabric defect types accurately.
+98%
Defect Detection Accuracy
3X
Faster Inspection Speed
-70%
Reduction in Quality Rework
