Kurdish Handwritten and Printed Text Recognition (OCR)

April 01, 2024
Ai
UKH Artificial Intelligence and Innovation Centre (AIIC) & Kurdish Academy
Lead Researcher / AI Project Director
1 min read

Highlights

Project Description

Handwritten Sorani recognition, printed text OCR, image preprocessing, line segmentation, and deep learning models such as TrOCR.

This project develops state-of-the-art optical character recognition pipelines specifically designed for both printed Kurdish texts and handwritten Central Kurdish (Sorani) script. It incorporates advanced preprocessing, morphological feature extraction, line and character segmentation, and transformer-based visual language models.

Technologies Used