A Novel Unified Approach to Deepfake Detection of Images

Lord Sen, Shyamapada Mukherjee

Abstract:

The advancements in the field of AI is increasingly
giving rise to various threats. One of the most prominent of
them is the synthesis and misuse of Deep Fakes. To sustain
trust in this digital age, detection and tagging of deepfakes is
very necessary. In this paper, a novel architecture for Deepfake
detection in images is presented. The architecture uses crossattention between spatial and frequency domain features along
with a blood detection module to classify an image as real or fake.
This paper aims to develop a unified architecture and provide
insights into each step. It is trained on two small datasets with
200 and 400 images respectively. On comparative analysis, our
model was better than the other possibilities. Further, there was
an increment in accuracy of 4.29% and 4.60% upon adding 200
images to the dataset. This shows that, if trained on a large
dataset and hyper-parameters optimized, the performance will
increase significantly.