Result Details
Task: Video Face Forgery Detection
Analysis Date: 2025-02-27 14:23:58.330393
Information:
Resolution: 480 x 854 | Codec: h264 | File Size: 1866.36KB | Bitrate: 484 kbps | Duration: 27.47 seconds | FPS: 25.153348873753647 |
Resolution: 480 x 854 | Codec: h264 | File Size: 1866.36KB | Bitrate: 484 kbps | Duration: 27.47 seconds | FPS: 25.153348873753647 |
Analytic Name: FTCN
Description: Deep learning based method for Video Face Forgery Detection that uses two stages: a fully temporal convolution network (FTCN) focusing on temporal feature extraction by reducing spatial kernel size to 1, and a Temporal Transformer network for long-term coherence exploration.
Reference:
https://arxiv.org/abs/2108.06693
Code Link:
https://github.com/yinglinzheng/FTCN
🧪 Fake Probability: 99.73%
â–²
99.73%
AI-generated Likelihood
Scope: Human face videos