Open Call project, 2023 Identifying stolen cultural heritage objects

Art theft and illegal trade has left many temples in Nepal depleted of their once rich furnishing of Hindu or Buddhist statues. The success rate of identifying stolen artefacts in international collections is highly limited by (a) the lack of systematic documentation and accessibility thereto as well as (b) the manpower needed to efficiently compare historic photographs of stolen artefacts with the sheer number of artefacts offered on the market and contained in collections. This project addresses these limitations by applying computer vision methods to identify similarities of objects with photographs from the DANAM database of Hindu monuments in Nepal.

The SSC developed a prototype for performing similarity search against the image data from the DANAM database. Image data is accessed directly from the university's media database HeidICON via its API. The tool can not only be used to identify stolen objects, but can also generate valueable insights for researchers working with cultural heritage objects.