Deep fake detection

Deep Fake Face Detection (딥러닝 위조 얼굴 검출)

1.LEE, Cheng-Han, et al. Maskgan: Towards diverse and interactive facial image manipulation. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2020. p. 5549-5558.

2.KARRAS, Tero, et al. Progressive growing of gans for improved quality, stability, and variation. arXiv preprint arXiv:1710.10196, 2017.

* 실험 환경

Dataset: 

  - CelebA-HQ1 이미지 10,000장

  - PGGAN2으로 생성된 이미지 10,000장

  - 14,000 training dataset, 4000 validation dataset, 2000 testing dataset

Test data:

  - Raw – 256X256

  - Blur – Kernel size 11, sigma 1.5 Gaussian Blur

  - Noise – Mean 0, sigma 1 Gaussian Noise

  - JPEG - Quality 75

* 논문 출처 (우수논문상 수상):

윤정민 이수빈 장해웅 정용주, 딥텍스처 특징에 기반한 위조 얼굴 검출(Fake Face Detection based on Deep Texture Features), 한국멀티미디어학회 춘계학술발표대회, 2021.05. 

가천대학교 AI소프트웨어학부, Computer Vision and Image Processing Lab