Transductive learning for reading handwritten tibetan manuscripts

Sivan Keret, Lior Wolf, Nachum Dershowitz, Eric Werner, Orna Almogi, Dorji Wangchuk

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

7 Scopus citations

Abstract

We examine the use case of performing handwritten character recognition (HCR) on a newly compiled collection of Tibetan historical documents, which presents multiple challenges, including inherent challenges such as image quality and the lack of word separation, and dataset challenges such as a lack of supervised training data. To tackle these challenges, we introduce an end-to-end unsupervised full-document HCR approach composed of unsupervised line segmentation and a convolutional recurrent neural network, trained using solely synthetic data. Various augmentations are applied to these synthesized images, and we compare the effect of each augmentation on the HCR results. Since we work on a collection of historical manuscripts, we can fit the model to the available test data. During training, our network has access to both the labeled synthetic training data and the unlabeled images of the test set, and we adapt and evaluate four different semi-supervised learning and domain adaptation approaches for transductive learning in HCR. We test our approach on a set of 167 images from the 'Kadam' collection, containing 829 lines. We show that correct data augmentation is crucial for the success of HCR trained solely on synthetic data and that using an effective transductive learning approach drastically improves results.

Original languageEnglish
Title of host publicationProceedings - 15th IAPR International Conference on Document Analysis and Recognition, ICDAR 2019
PublisherIEEE Computer Society
Pages214-221
Number of pages8
ISBN (Electronic)9781728128610
DOIs
StatePublished - Sep 2019
Event15th IAPR International Conference on Document Analysis and Recognition, ICDAR 2019 - Sydney, Australia
Duration: 20 Sep 201925 Sep 2019

Publication series

NameProceedings of the International Conference on Document Analysis and Recognition, ICDAR
ISSN (Print)1520-5363

Conference

Conference15th IAPR International Conference on Document Analysis and Recognition, ICDAR 2019
Country/TerritoryAustralia
CitySydney
Period20/09/1925/09/19

Keywords

  • CRNN
  • Domain Adaptation
  • Handwritten Recognition
  • Historical Document Analysis
  • Neural Network
  • Synthetic Data
  • Transductive Learning

Fingerprint

Dive into the research topics of 'Transductive learning for reading handwritten tibetan manuscripts'. Together they form a unique fingerprint.

Cite this