FACE SWAPPING & DEEP FAKE DETECTION

Abstract

The project entails the development of an AI-based Face Swapping & Deepfake Detection System that does two things, creates face-swapped outputs, and detects altered content very well. The faceswapping component employs deep learning-based models to capture facial characteristics of a source image and inlay them with desired level of realism onto the corresponding target facial region keeping the expressions and lighting conditions unaltered. The deepfake detection module employs transfers learning with pre trained models like Res Net or GANs trained on large datasets of authentic and fake images.