Deepfake Detection
Full-stack app that detects deepfake images with high accuracy using an EfficientNet model served through a FastAPI backend.
A full-stack application designed to detect deepfake images with high accuracy. The React-based frontend provides a user-friendly interface for uploading images and instantly viewing detection results. The FastAPI backend handles image processing and leverages a trained EfficientNet model to generate predictions, serving them via a REST API. Demonstrates seamless integration between modern frontend technologies and advanced deep learning models for real-time inference.
- EfficientNet classifierFine-tuned image classifier delivering high-accuracy fake vs real predictions.
- REST inference APIFastAPI endpoint accepts uploads and returns predictions in under a second.
- React uploaderDrag-and-drop interface with instant preview and result overlay.
Balancing model size against inference latency, and preventing false positives on compressed images.
Gained hands-on experience deploying deep-learning models behind a production-style API and streaming predictions to a React UI.
