The Scientific Document Analysis Group works on the recognition, analysis and enhancement of documents. We are particularly interested in the challenges posed by documents from scientific disciplines but also work on various other types of specialised documents, e.g.. historical or technical documents or written material captured from photographs, electronic whiteboards or tablet computers.
We work on optical images of documents as well as digitally born material - i.e. PDF or other formats generated directly by document formatting or production systems.
Our core activities centre on the recognition of formulae in the literature of mathematics and related scientific disciplines, the development of algorithms to recognise specialist symbols, diagrams and artifacts and syntactic and semantic approaches for the enhancement and verification of recognition results.
One aim of the work is to bridge the gap between literature and software by providing means to make the former more easily translatable into the later by automated means. The group therefore also investigates how intuitive mathematical concepts, as they often occur in mathematical writing, can be represented in symbolic computation and reasoning systems.
Document enhancement activities we are pursuing include enabling scientific documents for the visually impaired and metadata extraction and correction to improve search abilities on target documents.