Learning Algorithms of Form Structure for Bayesian Networks

Emilie Philippot (1) , Yolande Belaïd (1) , Abdel Belaïd (1)

1 READ - READ

In this paper, a new method is presented for the recognition of online forms filled manually by a digital-type clip. This writing process is not very restrictive but it is only sending electronic ink without the pre-printed form, which will require to undertake field recognition without context. To identify the form model of filled fields, we propose a method based on Bayesian networks. The networks use the conditional probabilities between fields in order to infer the real structure. We associate multiple Bayesian networks for different structures levels (i.e. sub-structures) and test different algorithms for form structure learning. The experiments were conducted on the basis of 3200 forms provided by the Actimage compagny, specialist in interactive writing processes. The first results show a recognition rate reaching more than 97%.

Mots clés: Baysian network; Form classification; On-line hand-writting

Domaines: Interface homme-machine [cs.HC]; Intelligence artificielle [cs.AI]; Traitement du texte et du document; Vision par ordinateur et reconnaissance de formes [cs.CV]

Emilie Philippot, Yolande Belaïd, Abdel Belaïd. Learning Algorithms of Form Structure for Bayesian Networks. International Conference on Image Processing - ICIP 2010, Sep 2010, Hong Kong, China. pp.2149-2152. ⟨inria-00526725⟩