Analysis and Detection of Deepfakes

Presented a pioneering deep learning method and introduced a unique approach using classical machine learning algorithms to extract and select classical features. This approach combines techniques from image processing, statistics, and computer vision. To evaluate its effectiveness, we conducted a thorough omnibus test under the null hypothesis, comparing the performance of multiple ML models across a range of statistical hypothesis testing frameworks.

Analysis and Detection of Deepfake (aka. image/video manipulation).