Text Recognition in Videos using a Recurrent Connectionist Approach

Khaoula Elagouni (1) , Christophe Garcia (2) , Franck Mamalet (1) , Pascale Sébillot (3)

1 Orange Labs R&D [Rennes]
2 imagine - Extraction de Caractéristiques et Identification
3 TEXMEX - Multimedia content-based indexing

Most OCR (Optical Character Recognition) systems developed to recognize texts embedded in multimedia documents segment the text into characters before recognizing them. In this paper, we propose a novel approach able to avoid any explicit character segmentation. Using a multi-scale scanning scheme, texts extracted from videos are first represented by sequences of learnt features. Obtained representations are then used to feed a connectionist recurrent model specifically designed to take into account dependencies between successive learnt features and to recognize texts. The proposed video OCR evaluated on a database of TV news videos achieves very high recognition rates. Experiments also demonstrate that, for our recognition task, learnt feature representations perform better than hand-crafted features.

Mots clés: Video text recognition; multi-scale image scanning; ConvNet; LSTM; CTC

Domaines: Multimédia [cs.MM]; Traitement du texte et du document; Traitement du signal et de l'image [eess.SP]; Traitement du signal et de l'image [eess.SP]

Khaoula Elagouni, Christophe Garcia, Franck Mamalet, Pascale Sébillot. Text Recognition in Videos using a Recurrent Connectionist Approach. 22th International Conference on Artificial Neural Networks, ICANN, Sep 2012, Lausanne, Switzerland. pp.172-179, ⟨10.1007/978-3-642-33266-1_22⟩. ⟨hal-00753906⟩