Seminar by Dr Sharad Seth
Recognizing Unsegmented Text by Symbolic Indirect Correlation
Dr Sharad Seth
Professor, University of Nebraska, Lincoln
Visiting Professor, Department of CSE, IIT Kanpur
Date: Fri, Feb 3, 2006
Time: 1530
Venue: CS 101Abstract:
I will describe a non-parametric approach to whole-word recognition that is being pursued in collaboration with Professors George Nagy (RPI) and Shashank Mehta (IITK). The approach builds two bipartite graphs that result from feature-level and lexical comparisons of the same word against a reference string which need not include the query word. The lexical graph preserves the relative order of edges in the feature graph corresponding to correctly recognized features. This observation leads to a subgraph-matching formulation of the recognition problem. Unlike parametric methods, such as hidden-Markov models (HMMs), the proposed approach does not require extensive training to estimate the model parameters.