Computational Historical Image Semantics
Associated since: January 2015
Prof. Dr. Bernhard Jussen
Prof. Dr. Alexander Mehler
While the fully automatic content analysis of images based on the model of automatic text analysis is likely to last longer than desideratum, text-image relationships offer a promising approach for the semi-automatic integration of texts and images into multimodal data types. This link in turn promises a series of innovations in the field of image- and language-oriented sciences. These are processes of knowledge in the humanities which start with the simultaneous evaluation of texts and images, in analogy to the fusion of arbitrary (for example linguistic or pictorial) information units characteristic of human information processing, which allows their mutual reference to content at any time. This is where the planned project comes in:
The project aims at the integration of symbolic and iconographic information units, which look at their interrelations as well as their compositional relations. Consequently, the project is located in the field of basic research in the humanities by asking about the possibilities and limits of multimodal information fusion and its technological reconstruction. The logical atoms of this approach form multimodal sign aggregates, which can be symbolically or iconographically manifested or contextualized.
In contrast to primarily or even purely computer-based text-image models, the project assumes the historicity of text-image relationships, according to which aspects of change or temporal variation affect not only the linguistic but also the iconographic side of sign constitution and reception. At the centre of the project is thus a historical science that is both text- and image-related, which is to be expanded into a computational historical image semantics, following the model of computational historical text semantics.
Authority Manager (https://authority.hucompute.org): Instance for managing users and groups and assigning access rights to resources. Access via REST-API.
OWLnotator (http://phi.owlnotator.hucompute.org/):Instance for managing ontologies and annotations. Access via REST-API.
Resource Manager (http://phi.resources.hucompute.org/help): Instance for managing resources. Access via REST-API.
Generic user interface: http://phi.resources.hucompute.org/gui/