St.

Vincent Pallotti College of Engineering and Technology

DEPARTMENT OF ARTIFICIAL INTELLIGENCE

Academic Year 2024-25

Project Seminar
On
“OBJECT CHARACTER
RECOGNITION”
Name of Guide: Project Members:
Mayank Charde (23010042)
Prof. Aparitosh Gahankari Tanay Makde(23010041)
Aditya Menon(23010062)
Nishant Tiwari(23010061)
Table of Content
• Introduction
• Problem Definition
• Literature Survey
• Comparative analysis of literature survey
• Tentative project flow(optional)
• Hardware & Software Requirements
• References
 Introduction
Object Character Recognition (OCR) is a technology that detects and extracts text from images or scanned
documents. It converts printed or handwritten characters into machine-readable text, enabling editing,
searching, and processing. Commonly used in digitizing documents, license plate recognition, and receipt
scanning.

Problem Statement
Extracting text from images and scanned documents manually is time-consuming and error-prone.
Traditional methods do not efficiently convert printed or handwritten text into a digital format, limiting
accessibility, searchability, and editability of important documents. There is a need for an automated solution
to accurately recognize and digitize text from various sources.

Objective
Develop an Optical Character Recognition (OCR) system that leverages machine learning to accurately
extract text from images and scanned documents. The system should efficiently convert printed or
handwritten text into digital form, enabling easier editing, searching, and storage of textual information.
 Literature Survey
Sr. no Title of paper Author name Name of conference / journal Year Findings

1 Optical Character Alaa Najmi Prachi Tiwari, Meetkumar 2024 An ml project OCR and NER for highly
Recognition and Named confidential docs focuses on accurately
Entity Recognition for Patel extracting text while ensuring sensitive data
Highly Confidential is securely identified and redacted.
Documents
2 General OCR Theory Anonymous Rajat Sharma, Adweteeya 2025 Model generalization, data and accuracy,
user-centric performance
Dwivedi
3 A Comprehensive Study of Ravi Raj A. Kos June 2022 This study examines the evolution of OCR
Optical Character technologies, highlighting their applications in
Recognition" document scanning and translation services. It
discusses the challenges related to accuracy and
processing time, emphasizing the need for
improved methodologies to enhance OCR
performance across various languages and scripts.
4 Handwritten Optical Character Jamshed Memon, Ahmed Khan Jan 2020 This systematic literature review analyzes research
Recognition (OCR) Maira Sami, Rizwan conducted on handwritten OCR from 2000 to 2018.
It synthesizes findings from 142 selected articles,
providing insights into the techniques used for
character recognition and identifying research gaps
for future exploration.

08-04-2025 4
 Algorithm-Techniques-Tools
Basic Setup (For Small-Scale OCR)

CPU : Intel i5 / AMD Ryzen 5 or higher

OS: Windows, Linux (Ubuntu), macOS Programming Language

Libraries & Tools : Flask==2.3.2

• Flask (2.3.2)
A lightweight web framework for building web applications and APIs in Python.
• Flask-CORS (4.0.0)
A Flask extension that enables Cross-Origin Resource Sharing (CORS)
support for your web app/API.
• Pytesseract (0.3.10)
A Python wrapper for Google’s Tesseract-OCR engine, used to extract text
from images.
• Pillow (10.0.0)
A powerful image processing library in Python; a modern fork of the PIL
(Python Imaging Library)
• Opencv-python-headless (4.8.0.74)
A headless version of OpenCV for image and video processing (no GUI
features included, ideal for servers).
• Numpy (1.24.3)
A core library for numerical computing in Python, often used for handling arrays
and matrices.
Completion Status
Screenshots of Work Done
 Thank you!

08-04-2025 8
 Any suggestions ?

08-04-2025 9