Bootstrapping Samples of Accidentals in Dense Piano Scores for CNN-Based Detection

Abstract : State-of-the-art Optical Music Recognition system often fails to process dense and damaged music scores, where many symbols can present complex segmentation problems. We propose to resolve these segmentation problems by using a CNN-based detector trained with few manually annotated data. A data augmentation bootstrapping method is used to accurately train a deep learning model to do the localization and classification of an accidental symbol associated with a note head, or the note head if there is no accidental. Using 5-fold cross-validation, we obtain an average of 98.5% localization with an IoU score over 0.5 and a classification accuracy of 99.2%.
Type de document :
Communication dans un congrès
2017 14th IAPR International Conference on Document Analysis and Recognition (ICDAR), Nov 2017, Kyoto, Japan. IEEE, 〈10.1109/ICDAR.2017.257〉
Liste complète des métadonnées

Littérature citée [3 références]  Voir  Masquer  Télécharger

https://hal.archives-ouvertes.fr/hal-01712213
Contributeur : Kwon-Young Choi <>
Soumis le : lundi 19 février 2018 - 11:42:04
Dernière modification le : mercredi 16 mai 2018 - 11:23:35
Document(s) archivé(s) le : lundi 7 mai 2018 - 12:30:59

Fichier

GREC_2017_camera_ready.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

Citation

Kwon-Young Choi, Bertrand Coüasnon, Yann Ricquebourg, Richard Zanibbi. Bootstrapping Samples of Accidentals in Dense Piano Scores for CNN-Based Detection. 2017 14th IAPR International Conference on Document Analysis and Recognition (ICDAR), Nov 2017, Kyoto, Japan. IEEE, 〈10.1109/ICDAR.2017.257〉. 〈hal-01712213〉

Partager

Métriques

Consultations de la notice

266

Téléchargements de fichiers

41