The Department of Information Technology (DIT) in the Kurdistan Region is responsible for overseeing the development of digital solutions for various governmental bodies. As part of their efforts to digitize and streamline public services, DIT sought a robust and high-performance AI solution capable of accurately extracting Kurdish and Arabic text from national ID cards. They partnered with Ovanya to build a custom OCR (Optical Character Recognition) system that could process ID card images, intelligently crop out unnecessary parts, and extract structured text data to be used in government services, particularly for the "My Account" initiative, which automates citizen data entry.
The Problem
Traditional OCR systems often struggle with accurately reading bilingual documents—especially when it comes to languages like Kurdish and Arabic that require context-specific handling of characters, alignment, and direction. DIT needed a solution that could not only extract text from scanned ID cards but also assign each piece of text to its correct field (e.g., name, date of birth, ID number) with high precision and reliability. Speed was also a concern—any delay in processing could bottleneck digital workflows, particularly in citizen services that involve high-volume data processing. DIT also required automation to streamline the data entry process for their "My Account" initiative, which aims to digitize citizen information for easier access and management.
The Solution
Ovanya developed a tailored API service powered by a custom-trained OCR model optimized for Kurdish and Arabic scripts. The API accepts images of national ID cards, automatically crops out irrelevant parts of the image, and performs text extraction using our high-accuracy model. Extracted text is then classified and mapped to predefined data fields, enabling easy integration with government services and databases, including the "My Account" initiative. By automating the data entry process, the solution eliminates manual entry errors, speeds up workflows, and ensures consistency across citizen records. Our solution was built on cutting-edge machine learning technologies and trained on real-world data provided by DIT, which allowed us to achieve high accuracy in both character recognition and field mapping. The final product was a lightweight, efficient, and secure API that met DIT’s performance and accuracy benchmarks. Ovanya continues to support and maintain the service, ensuring smooth operations and adaptability to future updates or document changes.
Other Projects
