Deep-fake-Forensics-Challenge

Shallow- and Deep- fake Image Manipulation Localization Using Deep Learning .

Shallow- and Deep-Fake Image Manipulation Localization

Overview

This repository contains the code and documentation for the CS Challenge on Shallow- and Deep-Fake Image Manipulation Localization Using Deep Learning. The challenge involves replicating and enhancing the results of a research paper, drafting a scientific paper, and formulating a startup idea.

You can read our paper from here: link.

Table of Contents

• Getting Started
• Usage
• Dataset
• Train/Val/Test

Getting Started

Follow these steps to get started with the project:

To begin, you will need to clone this GitHub repository :

git clone https://github.com/Goodnight77/Deep-fake-Forensics-Challenge.git

Navigate to the project directory:

cd Deep-fake-Forensics-Challenge

Install the required packages:

pip install -r requirements.txt

Usage

To run the project, use the following command:

python main.py --n_epoch 50 --n_bs 32 --train_dataframe /path/to/train_data.csv --val_dataframe /path/to/val_data.csv

Dataset

The whole dataset we used to train our model can be downloaded from kaggle using this link.

Train/Val/Test

The way (file paths) we split the datasets into train/val/test subsets can be downloaded link

Download saved model

link to an old saved model: https://drive.google.com/file/d/1-0ayZgLQpDqtt5NORYzYCNzGoaF6VIWY/view?usp=sharing