Abstract:Developing an Optical Character Recognition (OCR) system for handwritten texts is a challenging research problem. Handwritten text can be largely different even for the same piece of text since the writing style differs from person to person. On the other hand, for many regional languages, unavailability of datasets having a large quantity of varied images is a hindrance for the research advancements. This is especially true for Bangla, which is the most widely spoken language of Bangladesh and the second most widely spoken language of India. Only a few works have been proposed for the generation of handwritten Bangla basic characters and almost none related to the generation of handwritten Bangla compound characters. To this end, in this work, a method for the generation of synthetic handwritten Bangla compound characters is proposed to alleviate this data scarcity. A generative adversarial network (GAN) based model is developed for this purpose taking inspiration from the recent Auxiliary Classifier GAN (AC-GAN) model. A novel dataset partitioning scheme is also developed for handwritten character related tasks to improve the performance of the model. The quality of generated samples is evaluated in terms of the Fréchet Inception Distance (FID) metric. It is observed that the present model performs better in comparison to the basic AC-GAN architecture and also in comparison with some present GAN architectures. The sample dataset generated as a part of this work is publicly available at https://github.com/hachiro-2001/Bengali_Compound_Characters.

A semi-automatic annotation scheme for Bangla online mixed cursive handwriting samples

An analytic scheme for online handwritten Bangla cursive word recognition

Offline recognition of handwritten Bangla characters: an efficient two-stage approach

Unconstrained Bangla online handwriting recognition based on MLP and SVM

Recognition of Bangla Handwriting: Current Status and Future Directions

An end-to-end system for Bangla online handwriting recognition

Database generation and recognition of online handwritten Bangla characters

On-line handwriting recognition of Indian scripts-the first benchmark

ISIgraphy: A tool for online handwriting sample database generation

An Improved Scheme for Off-line Recognition of Handwritten Bangla Characters

Script independent online handwriting recognition

Bangla online handwriting recognition using recurrent neural network architecture

Online handwritten Bangla character recognition using HMM

Direction code based features for recognition of online handwritten characters of Bangla

Levenshtein distance metric based holistic handwritten word recognition

A hybrid model for end to end online handwriting recognition

A robust scheme for extraction of text lines from handwritten documents

Segmentation of Offline Handwritten Bengali Script

Generation of a synthetic handwritten Bangla compound character dataset using a modified conditional GAN architecture

Classification of Bangla Compound Characters Using a HOG-CNN Hybrid Model

On recognition of handwritten Bangla characters