On the Use of Attention Mechanism in a Seq2Seq based Approach for Off-line Handwritten Digit String Recognition - CNRS - Centre national de la recherche scientifique Accéder directement au contenu
Communication Dans Un Congrès Année : 2019

On the Use of Attention Mechanism in a Seq2Seq based Approach for Off-line Handwritten Digit String Recognition

Résumé

In this work, we investigate the use of the attention mechanism in deep learning for a better reading of handwritten digit strings in digitized images. The proposed recognition system built upon a CNN (Convolutional Neural Network) and two RNNs (Recurrent Neural Networks), acting as Encoder and Decoder and using the attention mechanism. We used a 1D mechanism for attention location with a "soft" alignment attention which has the peculiarity of having an easily calculable gradient and thus to integrate well with the network. Experimental results on data from ORAND-CAR A, ORAND-CAR B and CVL HDS databases compare favorably to other published methods.
Fichier principal
Vignette du fichier
Thiebault-ICDAR.pdf (1.02 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-02460896 , version 1 (30-01-2020)

Identifiants

  • HAL Id : hal-02460896 , version 1

Citer

T Lupinski, A. Belaid, A Kacem Echi. On the Use of Attention Mechanism in a Seq2Seq based Approach for Off-line Handwritten Digit String Recognition. ICDAR, Sep 2019, Sydney, Australia. ⟨hal-02460896⟩
38 Consultations
327 Téléchargements

Partager

Gmail Facebook X LinkedIn More