The deep learning model is trained on thousands of images handloom and power loom images of gamucha. Credit: Chiring Chandan, CC BY-SA 4.0
To check counterfeiting of the iconic gamucha towels of Assam, computer scientists have created an artificial intelligence tool that can quickly identify the authentic handloom fabric from machine-made replicas1.
The deep learning model is based on a dataset of 17,484 images featuring handloom and power loom gamucha fabrics. They used this dataset to train and evaluate six already existing deep learning architectures alongside their new, tailor-made model.
The new model demonstrated superior performance than established ones, especially in computational efficiency and adaptability, crucial for its practical deployment in the textile industry.
The technological innovation comes with social and economic benefits. Empowering handloom experts with an automated tool for product differentiation can safeguard cultural heritage and supports the livelihoods of traditional weavers. The integration of AI can also foster supply chain transparency, mitigating market fraud and enhancing consumer confidence in genuine handloom products, the researchers say.
The tool can help heritage textiles thrive in a digital landscape, they say.
