论文信息 - A Multi-stream Approach to Off-Line Handwritten Word Recognition

A Multi-stream Approach to Off-Line Handwritten Word Recognition

We present in this paper a new approach based on multi-stream hidden Markov models (HMM) for the recognition of off-line handwriting. Every word is presented by two HMM models: the first one is learned with features extracted from upper contour, the second with features extracted from lower contour. The combination of these two sources of information is studied using the multi-stream framework. We present experiment results obtained on a database composed of isolated words extracted from incoming mail documents.

Yousri Kessentini | Thierry Paquet | Abdelmajid Benhamadou

[1] Bernadette Dorizzi,et al. A Hidden Markov Models combination framework for handwriting recognition , 2003, Document Analysis and Recognition.

[2] Hervé Bourlard,et al. Multi-Stream Speech Recognition , 1996 .

[3] Jussi Kangasharju,et al. The use of meta-HMM in multistream HMM training for automatic speech recognition , 1998, ICSLP.

[4] Jr. G. Forney,et al. The viterbi algorithm , 1973 .

[5] George Saon,et al. Modèles markoviens uni- et bidimensionnels pour la reconnaissance de l'écriture manuscrite hors-ligne. (One and two-dimensional Markov models for off-line handwriting recognition) , 1997 .

[6] Samy Bengio,et al. Offline recognition of unconstrained handwritten texts using HMMs and statistical language models , 2004, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[7] Bernadette Dorizzi,et al. Strategies for combining on-line and off-line information in an on-line handwriting recognition system , 2001, Proceedings of Sixth International Conference on Document Analysis and Recognition.

[8] Hervé Bourlard,et al. Subband-based speech recognition , 1997, 1997 IEEE International Conference on Acoustics, Speech, and Signal Processing.