RSIS Repository Open-access research from RSIS International journals

International Journal of Research and Innovation in Applied Science (IJRIAS)

AI-Driven Model for Converting Modi Lipi Documents into the English Language

byMr. Anup Arun Govande

Published April 2, 2026  •  Vol. 11, Issue 3, pp. 328–334Open Access
DOI: 10.51584/IJRIAS.2026.11030031

Abstract

Modi Lipi is an old, cursive script used for centuries to write records in the Maratha Empire and neighbouring regions from approximately the 13th to the early 20th century. Today, hundreds of thousands of these documents are stuck in archives because very few people can still read them. To fix this, through this paper researcher created ModiAnuwad, an AI system model that automatically reads these handwritten scripts and translates them into English. The process works in five steps: it cleans up the document images, breaks them into individual characters, identifies them using a powerful neural network, translates the text using a "Transformer" model (similar to the tech behind ChatGPT), and then fixes any grammar mistakes.

Keywords: Modi Lipi; Historical document recognition; OCR, ANN

JournalInternational Journal of Research and Innovation in Applied Science (IJRIAS)
ISSN2454-6194
Volume / IssueVolume 11, Issue 3
Pages328–334
Publication dateApril 2, 2026
DOI10.51584/IJRIAS.2026.11030031
PublisherRSIS International
LicenseOpen Access

How to cite this article

Mr. Anup Arun Govande (2026). AI-Driven Model for Converting Modi Lipi Documents into the English Language. International Journal of Research and Innovation in Applied Science (IJRIAS), 11(3), 328-334. https://doi.org/10.51584/IJRIAS.2026.11030031

BibTeX

@article{Mr2026,
  title   = {AI-Driven Model for Converting Modi Lipi Documents into the English Language},
  author  = {Mr. Anup Arun Govande},
  journal = {International Journal of Research and Innovation in Applied Science (IJRIAS)},
  volume  = {11},
  number  = {3},
  pages   = {328--334},
  year    = {2026},
  doi     = {10.51584/IJRIAS.2026.11030031},
  publisher = {RSIS International}
}