Deepfake Studio

AI Detection & Generation Platform

Full-stack DeveloperAI SpecialistResearch Engineer
FPT University logoFPT University
Timeline

01/2025 - 04/2025

Duration

4 months

Role
Full-stack DeveloperAI SpecialistResearch Engineer
Technologies
FlaskPyTorchResNet50+9

Project Overview

DeepfakeStudio is a web application built with Flask that offers both deepfake detection and generation capabilities. The project focuses primarily on research and implementation of various deepfake detection models, while also providing a face-swapping generation feature. I served as Full-stack Developer & AI Specialist, responsible for deploying detection models to web application, developing face swapping functionality, and creating Shallow Neural Network model with ~82% accuracy.

Technologies & Tools

Flask
PyTorch
ResNet50
OpenCV
ONNX
MediaPipe
InsightFace
VidGear
JavaScript
HTML/CSS
Deep Learning
Computer Vision

Backend & AI

FlaskPyTorchJavaScriptDeep LearningComputer Vision

Frontend & UI

HTML/CSS

Challenges

Implementing real-time deepfake detection with high accuracy

Optimizing model performance for web deployment

Creating intuitive interface for complex AI functionalities

Balancing detection accuracy with processing speed

Key Learnings

Mastered deepfake detection algorithms and neural network architectures

Gained experience in deploying AI models to web applications

Learned advanced computer vision techniques for face manipulation

Developed skills in research-oriented software development

Project Gallery

Real face detection

Real face detection with confidence score

Fake face detection

Deepfake detection showing manipulation indicators

Interested in working together?

Let's discuss your next project and how I can help bring your ideas to life.

Get in touch