A language-neutral method for searching online handwritten notes is provided. The different algorithms contained in this method enable querying online multilingual handwritten documents with substrings of words rather than just whole words. More particularly, two approaches are presented - one based on partial Fréchet distance calculations and the other based on a pair hidden Markov models. The partial Fréchet distance is adapted from the traditional Fréchet distance concept to match a subcurve or prefix of a query word. The pair hidden Markov model used in the present application is adapted from pair hidden Markov models used in bioinformatics as generative models of local and global alignment of biological sequences.
申请公布号
WO2011021130(A1)
申请公布日期
2011.02.24
申请号
WO2010IB53606
申请日期
2010.08.10
申请人
INDIAN INSTITUTE OF SCIENCE;BHATTACHARYYA, CHIRANJIB;RAMASWAMY, SRIRAGHAVENDRA