Overview of Computer Vision-Based Tools in Search and Analysis of Historical Content

Thanks/Due to progress in image analysis, researchers in the Digital Humanities (DH) can now use sophisticated computer vision systems for their work. Deep learning models have led to the implementation of powerful search systems, which are also suitable for historical and artistic content. However, current systems often lack controllability for users and therefore lead to an unsatisfying user experience. This paper explores research in this domain and analyses the potential of several current search and analysis tools. Tools based on deep learning computer vision yield sometimes unexpected but highly promising results; however, the controllability for users needs to be improved.