Deepfake-detection-with-FSBI

arevpapikyan/Deepfake-detection-with-FSBI

"Deepfake Detection with Frequency-Enhanced Self-Blended Images" is my capstone project for the Bachelor of Science in Data Science at the American University of Armenia.

Python Stars: 0 Forks: 0 ML/AI

Summary

A capstone project extending the Frequency-Enhanced Self-Blended Images (FSBI) method for deepfake detection by exploring the impact of alternative color spaces. It implements a deep learning pipeline using EfficientNet-B4 to classify images enhanced with Discrete Wavelet Transform features, trained and evaluated on the FaceForensics++ and Celeb-DF-v2 datasets.

Similar Projects