DFG Research Training Group 2159 - Integrity and Collaboration in Dynamic Sensor Networks (i.c.sens) News and Events News from the Faculty
The Project SILKNOW, dedicated to the preservation of the European silk heritage, wins the European Heritage Award / Europa Nostra Award 2022

The Project SILKNOW, dedicated to the preservation of the European silk heritage, wins the European Heritage Award / Europa Nostra Award 2022

Classifying silk fabrics with AI: LUH's Institute for Photogrammetry and GeoInformation plays a key role in an international project.

Silk fabrics represent a highly endangered part of our cultural heritage. For centuries, this material has stood for craftsmanship, beauty and luxury, from tapestries to flags, sword sheaths and wedding dresses to traditional costumes. Throughout centuries, the silk trade facilitated the exchange of ideas and innovations. For example, punch cards were first used in jacquard silk looms long before modern computers were even invented. However, these artisanal weaving techniques are increasingly in danger of being forgotten. The project SILKNOW, funded by the EU’s Horizon 2020 programme, aimed to preserve and disseminate the heritage of silk production in Europe, also using state-of-the-art artificial intelligence (AI) methods. The project, which involved institutions from Spain, Italy, France, Slovenia and Poland in addition to Leibniz Universität Hannover (LUH), has now won the European Heritage Award / Europa Nostra Award in the category Research. This renowned award has been honouring outstanding achievements in the field of cultural heritage preservation for 20 years.

In the context of the project SILKNOW, which ran from 2018 to 2021, new methods based on the latest techniques from information and communication technology have been developed to preserve the intangible heritage of ancient weaving techniques. A multidisciplinary team collaborated to this end, encompassing fields such as computer science, image processing, text analysis, but also art history, 3D printing, textile production and conservation. The project was coordinated by the University of Valencia (Spain), and LUH was represented by the Institute of Photogrammetry and GeoInformation. "We have developed artificial intelligence methods to derive information about historical silk fabrics, e.g. about time and place of origin, from images of such fabrics," explains A/Prof. Franz Rottensteiner, the project leader at LUH. With the help of these methods, information about silk fabrics that was originally not available in digital data sets can be automatically derived from images. A second method, developed in Hanover on a similar methodological basis, is used to search for records based on the similarity of images. Mareike Dorozynski, a doctoral student who developed these methods at LUH, explains: "Using this search function, you can learn something about a silk fabric of which you only have an image. The function searches the database for the fabrics most similar to the given image, which provides you with access to the information associated with these fabrics."

From existing data, an artificial neural network learns, for example, what silk fabrics of a certain era look like, and then applies this learned model to other images of silk fabrics. The automatic classification can be used to supplement collections that contain incomplete or non-standardised data. For cultural historians, the standardized information and extended search functions will make it easier to apply specific searches across different collections in the future. For training purposes, a knowledge graph generated within the framework of SILKNOW with almost 40,000 entries on artefacts made of silk with pictures and additional information could be used. As a result, information, for example about the place and time of production of a fabric, can now be automatically derived from images.

In addition, the project has created a "virtual loom" in which various weaving techniques, often known only to craftsmen, are visualised and preserved for future generations. Another result of SILKNOW is a multilingual thesaurus that contributes to the standardisation of terms in the analysis and description of silk fabrics and makes it possible to use terms and phrases in several languages for search queries.

The European Heritage Awards / Europa Nostra Awards jury''s citation states: "SILKNOW has created an innovative system to facilitate the transfer of the knowledge of silk weaving. This project represents an important example of how crafts, and therefore intangible heritage, can be linked to digital tools and how these tools can be used to democratize access to technical knowledge."

The winners will be honoured at a ceremony in Prague on 26 September in the presence of Mariya Gabriel, European Commissioner for Innovation, Research, Culture, Education and Youth. A total of 30 outstanding achievements from 18 European countries will be honoured. From 11 August to 11 September, it will also be possible to vote for the project to get the Public Choice Award using the following link: https://vote.europanostra.org  

 

More information about SILKNOW: https://silknow.eu

More information about the European Heritage Award / Europa Nostra Award: www.europeanheritageawards.eu

https://silknow.eu