Two-Stream Neural Networks for Tampered Face Detection
computer vision and pattern recognition, 2017.
We propose a two-stream network for face tampering detection. We train GoogLeNet to detect tampering artifacts in a face classification stream, and train a patch based triplet network to leverage features capturing local noise residuals and camera characteristics as a second stream. In addition, we use two different online face swaping ap...More
PPT (Upload PPT)