Matching the Stitch

Dr Xavier Aure and Dr George Jenkinson have been awarded a grant from UWE’s Arts and Humanities Research Council Impact Acceleration funds for a project to create a new digital toolkit which will use deep learning techniques to segment and classify stitched textures from photographs of fragile and important museum textile artefacts, providing an easy-to-use means to reverse engineer digital production processes capable of producing handling replicas with true-to-reality textures.

Advances in the ability of artificial intelligence models to classify images and development of inexpensive production methods demonstrate a level of technological readiness appropriate to making this digital toolkit a reality. The digital files output will be ready for production on a variety of computer controlled production techniques, including but not limited to, machine embroidery, SLS 3D printing, FDM 3D printing, laser engraving, and CNC milling. Part of the work will include cataloguing a database of different stitch types and developing a model that can classify them.

This project aims to significantly advance knowledge about historic textiles and will be of benefit to archivists, curators, and the general public.