CNN-based multilingual handwritten numeral recognition: A fusion-free approach

作者：

Highlights：

• Handwritten numeral classification of 8 different scripts/languages is proposed.

• Script free handwritten numeral recognition that abolishes multilingual concept.

• This class-based system solves multiscript numeral recognition as 10 class problem.

• Proposed system is flexible enough to extend for any new language readily.

• Has practical utility for sorting the postal cards with multilingual zip codes.

摘要

•Handwritten numeral classification of 8 different scripts/languages is proposed.•Script free handwritten numeral recognition that abolishes multilingual concept.•This class-based system solves multiscript numeral recognition as 10 class problem.•Proposed system is flexible enough to extend for any new language readily.•Has practical utility for sorting the postal cards with multilingual zip codes.

论文关键词：Convolution neural network,Handwritten,Multilingual,Numeral Recognition,Script Independent

论文评审过程：Received 11 October 2019, Revised 19 June 2020, Accepted 19 July 2020, Available online 5 August 2020, Version of Record 14 August 2020.

论文官网地址：https://doi.org/10.1016/j.eswa.2020.113784