AI & Computer Vision
Face Swapper
Overview
A modern, high-quality face swapping and photo enhancement web application powered by Streamlit, InsightFace, and GFPGAN. Features automatic facial landmark analysis, high-fidelity source-to-target swapping, and deep learning facial restoration with a sleek dark neon interface.
Project Highlights
- Deep Learning Face Swapping & Restoration: InsightFace analysis and GFPGAN face restoration for ultra-crisp swapped faces.
- Local & GPU-Accelerated Processing: Runs locally with ONNX Runtime and CUDA GPU acceleration with one-click download.
- Tech Stack: Python, Streamlit, InsightFace, GFPGAN, ONNX Runtime, OpenCV.
- Source & Live App: GitHub Repository
The challenge
Optimizing deep learning inference pipelines with ONNX Runtime, synchronizing 512-dim facial landmark embeddings in real-time, and integrating GFPGAN blind face restoration to produce photorealistic results at near-instant speeds.
Project Walkthrough & Explanation Video
To start a project, collaborate, or discuss opportunities, this is the best place to begin
GET IN TOUCH
Tell me a bit about what you're working on or opportunities you have — I'll get back to you promptly