The face deepfake detection challenge

Guarnera L., Giudice O., Guarnera F., Ortis A., Puglisi G., Paratore A., Bui L. M. Q., Fontani M., Coccomini D. A., Caldelli R., Falchi F., Gennaro C., Messina N., Amato G., Perelli G., Concas Sara, Cuccu C., Orru G., Marcialis G. L., Battiato S.

Deepfake detection  deepfake detection  Computer Vision and Pattern Recognition  deepfake reconstruction  Computer Graphics and Computer-Aided Design  Deep Learning  Nuclear Medicine and imaging  deep learning  discrete cosine transform  Transformer networks  Deepfake challenge  Electrical and Electronic Engineering  Deepfake reconstruction  transformer networks  deepfake challenge  Radiology  Discrete cosine transform 

Multimedia data manipulation and forgery has never been easier than today, thanks to the power of Artificial Intelligence (AI). AI-generated fake content, commonly called Deepfakes, have been raising new issues and concerns, but also new challenges for the research community. The Deepfake detection task has become widely addressed, but unfortunately, approaches in the literature suffer from generalization issues. In this paper, the Face Deepfake Detection and Reconstruction Challenge is described. Two different tasks were proposed to the participants: (i) creating a Deepfake detector capable of working in an "in the wild" scenario; (ii) creating a method capable of reconstructing original images from Deepfakes. Real images from CelebA and FFHQ and Deepfake images created by StarGAN, StarGAN-v2, StyleGAN, StyleGAN2, AttGAN and GDWCT were collected for the competition. The winning teams were chosen with respect to the highest classification accuracy value (Task I) and "minimum average distance to Manhattan" (Task II). Deep Learning algorithms, particularly those based on the EfficientNet architecture, achieved the best results in Task I. No winners were proclaimed for Task II. A detailed discussion of teams' proposed methods with corresponding ranking is presented in this paper.

Source: JOURNAL OF IMAGING 8 (2022). doi:10.3390/jimaging8100263

Metrics