Silk textiles have become a seriously endangered cultural heritage. For hundreds of years, silk has been synonymous with craftsmanship, beauty and luxury - used in contexts such as tapestries and flags, sword sheathes and wedding gowns, as well as traditional costumes. Silk trade has for centuries enabled the exchange of ideas and innovations, seeing punched cards used for the first time in Jacquard silk looms, long before modern computers were even imagined. However, these artisanal weaving techniques are at ever more risk of disappearing. With funding provided by the EU Horizon 2020 programme, SILKNOW aims to preserve and showcase the history of silk production in Europe using the most advanced methods of artificial intelligence. In addition to Leibniz University Hannover (LUH), institutions in Spain, Italy, France, Slovenia and Poland were involved in the international collaboration project, which has now also been awarded the European Heritage Award / European Nostra Award in the research category. For more than 20 years, the prestigious award has celebrated remarkable work in the field of cultural heritage.
During the lifespan of the project, which ran from 2018 to 2021, new methods based on state-of-the-art procedures from IT and communications technology were developed to preserve the intangible heritage of ancient weaving techniques. To achieve this, a multidisciplinary team has collaborated, comprising areas such as IT, image processing and text analysis, as well as artificial intelligence, 3D printing, textile fabrication and textile conservation. The project was coordinated by Universitat de Valencia (Spain), with LUH represented by the Institute of Photogrammetry and Geoinformation (Faculty of Civil Engineering and Geodetic Science). "We utilise methods from artificial intelligence to derive information about the depicted textiles - such as time and place of manufacture - from images of historical silk" explains project manager apl. Prof. Dr. Franz Rottensteiner. Using these methods, information about silk textiles that was not originally available in the digital datasets are automatically deduced from images. Similarly, another method developed in Hannover enables the search for datasets based on similarity of images. Mareike Dorozynski, who developed this method as a doctoral candidate at LUH, explains: "Using the search function it is possible to find out about a silk textile that you only have an image of. A search of the database is performed to find the textiles most similar to the image, which then gives access to the information linked to these textiles."
Using existing data, an artificial neural network learns, for example, the visual characteristics of silk textiles from a certain era and then applies this model for other images of silk textiles. Automatic classification can be used to complete collections that contain incomplete or non-standardised data. Unified information and expansive search functions will make it easier for cultural historians to search various collections in the future. Within the scope of SILKNOW, a knowledge graph featuring nearly 40,000 entries on artefacts made of silk with images and further information could be used for training. It is now possible to automatically derive information such as the time and place of production from images.
The project also produced a "Virtual Loom", in which different weaving techniques - often only known to those in the trade - are visualised and preserved for future generations. A further result of SILKNOW includes a multilingual thesaurus, which contributes to standardised use of terms when analysing and describing silk textiles, and enables terms and phrases to be searched in different languages.
According to the awards jury of the European Heritage Awards / Europa Nostra Awards "SILKNOW has developed an innovative system, which facilities the transfer of knowledge on silk weaving. This project is a significant example of how craftsmanship and the associated intangible cultural heritage can be linked using digital resources and how resources can be used to democratise access to technical knowledge."
The winners will receive their award from Mariya Gabriel - EU Commissioner for Innovation, Research, Culture, Education and Youth - at a ceremony in Prague on 26 September, where a total of 30 outstanding achievements from 18 European countries will be recognised. From 11 August to 11 September, members of the public can vote for the project that they think should receive the Public Choice Award via the link below: https://vote.europanostra.org
SILKNOW is also delighted about a further accolade. The project was included in the European Commission''s Innovation Radar - an initiative by the European Commission to identify innovation with great potential in EU-funded research and innovation projects. The platform aids the visibility and accessibility of information on innovation in these valuable projects: https://www.innoradar.eu/innovation/42892
Further information on SILKNOW: https://silknow.eu
Further information on the European Heritage Award / Europa Nostra Award: www.europeanheritageawards.eu
Note to editors
For further information, please contact apl. Prof. Dr. Franz Rottensteiner, Institute of Photogrammetry and Geoinformation (Tel. +49 762 - 3893, Email rottensteiner@ipi.uni-hannover.de).
