Arabic OCR System

The document outlines a proposed project for developing an Arabic Optical Character Recognition (OCR) system using YOLOv8, aimed at detecting text in images and videos. The project includes objectives such as leveraging Roboflow for dataset annotation, implementing real-time detection with OpenCV, and evaluating model performance. The expected outcomes include a trained YOLOv8 model, a real-time processing system, and a comparative performance analysis of different YOLOv8 versions.

Uploaded by

kmarry1990

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

57 views3 pages

Arabic OCR System

Uploaded by

kmarry1990

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 3

COMSATS UNIVERISTY ISLAMABAD, WAH CAMPUS

TOPICS IN COMPUTER SCIENCE- II

Proposed Projects: Alif-AI

Submitted by: FA21-BCS-049 (Maryam Khan)
FA21-BCS-085(Hijab)
FA21-BCS-072 (Neha Siddique)
Section: 8-A
Submitted to: Dr Saeed Ur Rehman
Dated: 25/02/2025
Title:

Arabic OCR System using YOLOv8

Introduction:

 Arabic Optical Character Recognition (OCR) plays a crucial role in digitizing and
processing Arabic text from images and videos. Traditional OCR techniques struggle
with handwritten and complex Arabic scripts due to varying fonts, styles, and
orientations. This project aims to develop a deep learning-based OCR system using
YOLOv8 for text detection in both images and video frames, leveraging Roboflow
for dataset annotation and training.

Objectives:

 To develop an Arabic OCR system capable of detecting text in images and video
frames.
 To use YOLOv8 for accurate text detection.
 To integrate Roboflow for dataset annotation and augmentation.
 To implement real-time text detection in videos using OpenCV.
 To evaluate model performance using accuracy, recall, precision, and F1-score.
 To create a lightweight and efficient OCR system suitable for real-world applications.

Methodology:

 Collect or generate an Arabic text dataset containing both printed and handwritten
text.
 Data Annotation Using Roboflow
 Model Training with YOLOv8
 Develop a Python-based system.
 Test for accuracy, precision, and recall.

Expected Outcomes:

 An Arabic text detection model that works on both images and videos.
 A trained YOLOv8 model capable of detecting Arabic text with high accuracy.
 A real-time video processing system for Arabic text recognition.
 A labeled dataset in YOLO format for further OCR research.
 A comparative performance analysis of different YOLOv8 versions.
Tools & Technologies:

 Programming Language: Python

 Deep Learning Framework: PyTorch
 YOLO Model: Ultralytics YOLOv8
 Dataset Annotation: Roboflow
 Computer Vision: OpenCV

Timeline:
Task Duration
Data Collection & Preprocessing 2 Weeks
Model Training & Testing 4 Weeks
OCR Integration 3 Weeks
UI Development & Final Testing 4 Weeks
Report Writing & Presentation 2 Weeks

Conclusion:

This project aims to develop an efficient Arabic OCR system using YOLOv8 for text
detection in images and videos. By utilizing Roboflow for dataset annotation and Ultralytics
YOLO for model training, the system will achieve high-speed and accurate text detection.
The developed solution will have potential applications in document digitization,
automated translation, and surveillance systems.