Deep-fake: An Emerging AI technology
一种新兴的人工智能技术
Course Syllabus 课程章节
1This course introduces the definition of deep-fake, lists several applications and different databases of fake faces, such as RefaceAI which is an AI-driven face-swapping application, mobile app Doublicat, Zao Chinese app, deepfakes web, etc. The deep-fake is a form of artificial intelligence combining the terms ‘deep learning’ and ‘fake’. in simple terms, it’s falsified videos made utilizing deep learning. Moreover, the course will explain how to generate a fake video, how to detect a fake video, and the future of deep-fake. Several deep learning models, such as DCGAN, PGGAN, StyleGAN are introduced to generate fake face images and videos. The key clues to detect fake videos cover several aspects including spatial features, temporal features, temporal-spatial features, features of external imaging environment, human biological signals, etc. The contents will also show several approaches to detect fake videos.