Recognition of Urdu Alphabets using deep learning

Several research studies have been carried out for recognition of Chinese, Japanese, English, and other language's text recognition from an image. Unfortunately, the research regarding Urdu text recognition from images is still immature due to cursive, variable and overlapping nature of Urdu writing

2025-06-28 16:34:45 - Adil Khan

Project Title

Recognition of Urdu Alphabets using deep learning

Project Area of Specialization Artificial IntelligenceProject Summary

Several research studies have been carried out for recognition of Chinese, Japanese, English, and other language's text recognition from an image. Unfortunately, the research regarding Urdu text recognition from images is still immature due to cursive, variable and overlapping nature of Urdu writing and different writing styles. In this project, we intend to work in this dimension using deep learning. The successful completion of the proposed project will facilitate the Urdu publishing in Pakistan by recognizing the Urdu text in an automated way and thus, bypassing contemmporary, erronous, and slow Urdu typewriting process and consequently decreasing the publishing cost. Additionally, it will reduce the  memory requirement. For this purpose, we intend to employ appropriate image acquisition, pre-processing, RGB to gray-scale conversion, Skew Correction, binarization, noise reduction, thinning, image features, or/and segmentation to read Urdu language from the image.

Project Objectives

1. Development of an Urdu character recognition system.

2. Publishing a research paper based upon our findings

Project Implementation Method

Recognition of Urdu Alphabets using deep learning _1582918164.png

Benefits of the Project

Most of the Urdu stuff already present in the ancient books is in form of images which consume a lot of memory s. If these images are converted into documents (having Urdu alphabets ASCII, then this will consume less memory space to transfer or to read online. Additionally, the benefits of the project are as follows:

Automated assistance to Urdu publishers in Pakistan.   
Cheaper publishing of Urdu publishing because of skipping the serviced of Urdu typewriter.

Technical Details of Final Deliverable

The final deliverable will be an application software which will need a scanned image having
various handwritten Urdu alphabets. The output of the application will be the sequence of
the Urdu alphabets in the image.

Final Deliverable of the Project Software SystemType of Industry IT Technologies Artificial Intelligence(AI)Sustainable Development Goals Quality Education, Industry, Innovation and InfrastructureRequired Resources
Item Name Type No. of Units Per Unit Cost (in Rs) Total (in Rs)
Total in (Rs) 35000
Scanner Equipment12500025000
Digital Handwriting Pad Equipment11000010000

More Posts