| Title | Deepfake Detection using a Two-Stream Capsule Network |
| Publication Type | Conference Paper |
| Year of Publication | 2021 |
| Authors | Joseph, Zane, Nyirenda, Clement |
| Conference Name | 2021 IST-Africa Conference (IST-Africa) |
| Date Published | may |
| Keywords | Analytical models, Capsule networks, Computational modeling, Computer architecture, convolutional neural networks, Deep Learning, DeepFake, deepfake detection, Error Level Analysis, face tampering, human factors, Metrics, Neural networks, Optimization, particle swarm optimization, pubcrawl, resilience, Resiliency, Scalability, Training |
| Abstract | This paper aims to address the problem of Deepfake Detection using a Two-Stream Capsule Network. First we review methods used to create Deepfake content, as well as methods proposed in the literature to detect such Deepfake content. We then propose a novel architecture to detect Deepfakes, which consists of a two-stream Capsule network running in parallel that takes in both RGB images/frames as well as Error Level Analysis images. Results show that the proposed approach exhibits the detection accuracy of 73.39 % and 57.45 % for the Deepfake Detection Challenge (DFDC) and the Celeb-DF datasets respectively. These results are, however, from a preliminary implementation of the proposed approach. As part of future work, population-based optimization techniques such as Particle Swarm Optimization (PSO) will be used to tune the hyper parameters for better performance. |
| Citation Key | joseph_deepfake_2021 |
Deepfake Detection using a Two-Stream Capsule Network