Passport Verification System Using OCR and Machine Learning (HOG + SVM)

This project presents a Machine Learning–based system designed to analyze passport images and extract textual information using Optical Character Recognition (OCR), along with detecting whether a passport image is real or potentially fake using image feature extraction techniques.

The system allows users to upload a passport image through an interactive interface, processes the image, extracts important text fields, and performs classification using a trained machine learning model.

Project features include:

Uploading passport images through a simple interface

Image preprocessing for better recognition accuracy

Extracting text using OCR (Tesseract)

Feature extraction using Histogram of Oriented Gradients (HOG)

Passport classification using an SVM model

Displaying extracted results clearly to the user

Technologies and libraries used:

Python

OpenCV

pytesseract (OCR)

scikit-learn

matplotlib

This system can be extended and applied in real-world scenarios such as:

Identity verification systems

Airport security automation

Government document validation

Intelligent document processing solutions

The project demonstrates practical implementation of Computer Vision and Machine Learning techniques