N. Bergboer & E. Postma & J. van den Herik
The increasing availability of digital historical visual data opens up new opportunities for cultural-heritage research. For instance, modern artificial-intelligence techniques enable historians to search for visual objects, such as persons and other figurative entities, in large historical image databases. Specialised techniques may offer the ability to detect automatically particular objects from a given class of objects. This paper presents our work on the development of accurate object-detection techniques that rely on contextual cues. We provide insight into the results of our technique and discuss how cultural-heritage researchers can benefit from intelligent object-detection techniques. From these results we may conclude that in the years to come, cultural-heritage research will change its research approaches significantly.
|