DEEPFAKE DETECTION SYSTEM

In this day and age there are many technologies that help us in daily life but there are some malicious people that take advantage of such technologies to cause harm to others one such technology is DeepFake's.  By using these DeepFake's some malicious users can and will discriminat

Project Title

DEEPFAKE DETECTION SYSTEM

Project Area of Specialization

Artificial Intelligence

Project Summary

In this day and age there are many technologies that help us in daily life but there are some malicious people that take advantage of such technologies to cause harm to others one such technology is DeepFake's. 

By using these DeepFake's some malicious users can and will discriminate people depicting of some form of wrongdoing or harmful action to others. To prevent such harm, we have  created a system that is capable to detect these deepfake images such that innocent people do not come under any threats.
 

Project Objectives

The aims and objectives of the system is to protect the identity of a person and are as follows: 

1. To help the users to save their data.
2. To help users to secure their identities/personal information from being misused. In short by using this application the user data and their personal information is secured by any random person.
3. To provide a secure and user-friendly environment where their data is secure from any unknown person. 
4. To provide a platform that can detect deep faked images and possibly videos and audios in the future.
5. To provide a platform for users to prevent misuse of their identity. 
6. To provide a platform for users and end-users to happily detect deep faked images to their satisfaction.
7. In the given points we have discussed what kind of features we are going to provide in our project and how it is going to provide a safe environment to the user.

The main goal of our project is to prevent misuse of anyone’s identity.

Project Implementation Method

The development process includes the system design in which we first designed the model or you can say simple structure that how it will work. After designing the system model, we trained the Artificial Intelligence (AI) model and then designed the front-end layout which is the user interface of the system. The layout includes all the UI that are in this system. The work on backend is done by using Python language on Visual Studio Code and at last the front-end and back-end is integrated with the trained AI model.

This system is working on the following architecture:

1. Our web-based application has a login and Sign up system.
2. After login, the main home page will be displayed to the user.
3. Now users can upload image for the detection process.
4. User can edit their information and also check user account log.
5. The main process of this system is to detect deepfake images.
6. User can detect their personal image or any image to check whether it's deep faked or not.

In addition, we also have used a basic MVC architecture with Flask mini-framework to further complement the Web Application.

We have used the Convolutional Neural Network (CNN) algorithm in the training of our DL model. We have also used the MesoNet which is the architecture of convolutional neural network in the DL model for the detection process.

We have also used MySQL & SQLServer to store user information( i.e.,Name, Age, DOB, etc...) on the system.
 

Benefits of the Project

They are as follows:

1. The project targets the local populace for its purposes(Detection of real or fake images) as well as the international one.
2. The project is a steping stone for any potential research in this technology in our country.
3. The project uses a custom made DL model that can be further trained and enhanced by simply retraining using any additional data(User provided images) that the users allow for the system to use.
4. The project is trained to detect DeepFakes on both the international and local stages so it can also detect on international faces as well. 
5. The project also uses a very simple interface to not cause stress to any low-end or high-end device.

Technical Details of Final Deliverable

The Final output of the project is to provide the user the satisfaction of the image to be either a real image or a fake user/AI generated image.

The image must be given in a .png, .jpeg or .jpg format for the system to accept and perform the detection process.

The user given image will than then be checked using the Systems trained DL model.

Finally, the given image will then be displayed in a neat fashion along side the result ( Real or Fake ) with an Accuracy rating (i.e., 95.654% etc...).

Final Deliverable of the Project

Software System

Core Industry

IT

Other Industries

Education , Legal , Telecommunication

Core Technology

Artificial Intelligence(AI)

Other Technologies

Augmented & Virtual Reality

Sustainable Development Goals

Good Health and Well-Being for People, Peace and Justice Strong Institutions

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com