Abstract
This paper addresses the indexing and retrieval of mathematical symbols from digitized documents. The proposed approach exploits Shape Contexts (SC) to describe the shape of mathematical symbols. Indexed symbols are represented with a vector space-based method that is grounded on SC clustering. We explore the use of the Self Organizing Map (SOM) to perform the clustering and we compare several approaches to compute the SCs. The retrieval performance are measured on a large collection of mathematical symbols gathered from the widely used INFTY database. © Springer-Verlag 2009.
Cite
CITATION STYLE
Marinai, S., Miotti, B., & Soda, G. (2009). Mathematical symbol indexing. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 5883 LNAI, pp. 102–111). https://doi.org/10.1007/978-3-642-10291-2_11
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
