In this paper we described an approach for automatic, template-based citation metadata extraction from scientific literature, as well as their visualization. The extraction approach assumes PDF format of file and IEEE reference writing standard. It is based on formally defined templates in form of regular expressions which are utilized to implement finite state machine for metadata extraction. After relations between references are extracted and stored in adequate data structures, their visualization using graphs and treemaps is performed. Graphs and treemaps proved to be very efficient and compact techniques for visualization of citation information, particularly effective to emphasize the most cited papers and authors in specific field of science. Finally, we demonstrated satisfied test results and discussed future plans for improvement of the approach.
bibliographies, finite state machines, information retrieval, information visualization