Image Forgery Detection using CASIA 2.0 Dataset

A deep learning-based image tampering detection project trained using the CASIA 2.0 Image Tampering Detection Dataset.

Overview

This project detects whether an image is:

• Real (Authentic)
• Fake (Tampered)

The model was trained using deep learning techniques for image forgery detection.

Features

• Detects tampered (fake) images
• Classifies images as Real or Fake
• Deep learning-based prediction
• Trained using CASIA 2.0 dataset
• Easy model retraining using train.py

Dataset Used

CASIA 2.0 Image Tampering Detection Dataset

Dataset contains:

• Authentic (Real) Images
• Tampered (Fake) Images

Technologies Used

• Python
• TensorFlow
• Keras
• OpenCV
• NumPy
• Flask

Project Structure

project-folder/
│── app.py
│── train.py
│── model/
│    └── forgery_model.h5

Model File

The trained model file (forgery_model.h5) is not uploaded because of GitHub file size limitations.

You can train the model again using:

python train.py

Or place the trained .h5 model file inside:

model/

Installation

pip install -r requirements.txt

Run Project

python app.py

Author

Bindu