Abstract
This paper explores synthetic image generators in dataset preparation to train models that allow human handwritten character recognition. We examined the most popular deep neural network architectures and presented a method based on autoencoder architecture and a schematic character generator. As a comparative model, we used a classifier trained on the whole NIST set of handwritten letters from the Latin alphabet. Our experiments showed that the 80% synthetic images in the training dataset achieved very high model accuracy, almost the same level as the 100% handwritten images in the training dataset. Our results prove that we can reduce the costs of creating, gathering, and describing human handwritten datasets five times over – with only a 5% loss in accuracy. Our method appears to be beneficial for a part of the training process and avoids unnecessary manual annotation work.
Research was funded by the Centre for Priority Research Area Artificial Intelligence and Robotics of Warsaw University of Technology within the Excellence Initiative: Research University (IDUB) programme (grant no 1820/27/Z01/POB2/2021).
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
Our source code for the handwritten character generator, schema examples are in: https://github.com/grant-TraDA.
References
Aaref, A.M.: English character recognition algorithm by improving the weights of MLP neural network with dragonfly algorithm. Period. Eng. Nat. Sci. 9, 616–628 (2021)
Alonso, E., Moysset, B., Messina, R.: Adversarial generation of handwritten text images conditioned on sequences. In: ICDAR (2019)
T.E. de Campos, B.B.: Chars74k dataset (2009)
Chaudhuri, A., et al.: Optical Character Recognition Systems for Different Languages with Soft Computing. Springer, Heidelberg (2017). https://doi.org/10.1007/978-3-319-50252-6
Chen, Q.: HIT-OR3C dataset. corpus for Chinese characters (2011)
Geetha, R.E.A.: Effective offline handwritten text recognition model based on a sequence-to-sequence approach with CNN-RNN networks. Neural Comput. Appl. 33, 10923–10934 (2021)
Ghosh, D., Shivaprasad, A.: An analytic approach for generation of artificial hand-printed character database from given generative models. Pattern Recognit 32, 907–920 (1999)
Jaderberg, M., Simonyan, K., Vedaldi, A., Zisserman, A.: Synthetic data and artificial neural networks for natural scene text recognition (2014)
Krishnan, P., Jawahar, C.V.: Generating synthetic data for text recognition (2016)
LeCun, Y.: The mnist database of handwritten digits (2019)
Mouchère, H.: Crohme: Competition on recognition of online handwritten mathematical expressions (2014)
NIST: Nist handprinted forms and characters - nist special database 19 (2019)
Santosh, K.: Devanagari character dataset (2011)
Simonyan, K., Zisserman, A.: Very deep convolutional networks for large-scale image recognition. In: ICLR (2015)
Slimane, F.: Arabic printed text image database (2009)
Taufique, M., et al.: Handwritten Bangla character recognition using inception convolutional neural network. Int. J. Comput. Appl. 181, 48–59 (2018)
Tensmeyer, C., Wigington, C.: Training full-page handwritten text recognition models without annotated line breaks. In: ICDAR (2019)
Veit, A., Matera, T., Neumann, L., Matas, J., Belongie, S.: Coco-text: dataset and benchmark for text detection and recognition in natural images. In: ICDAR (2017)
Zhang, J.: C-Bézier curves and surfaces. Graphical Models and Image Processing (1999)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2021 Springer Nature Switzerland AG
About this paper
Cite this paper
Wróblewska, A., Chechliński, B., Sysko-Romańczuk, S., Seweryn, K. (2021). How Much Do Synthetic Datasets Matter in Handwritten Text Recognition?. In: Mantoro, T., Lee, M., Ayu, M.A., Wong, K.W., Hidayanto, A.N. (eds) Neural Information Processing. ICONIP 2021. Lecture Notes in Computer Science(), vol 13110. Springer, Cham. https://doi.org/10.1007/978-3-030-92238-2_12
Download citation
DOI: https://doi.org/10.1007/978-3-030-92238-2_12
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-92237-5
Online ISBN: 978-3-030-92238-2
eBook Packages: Computer ScienceComputer Science (R0)