Check Out How To Track Fraud on Video

Liputan6.com, Jakarta- It used to be that only a deep-pocketed Hollywood film company with a team of skilled artists and technicians could make video deepfake – forgery realistic, which shows someone doing and saying things they never did or said.

Not anymore. Quoted from theconversation.com, software available in cyberspace for free allows anyone who has a computer to make videos false convincing in an instant.

Whether used for personal grudges, to humiliate celebrities, or to sway public opinion, deepfakes today refute the ancient axiom of seeing then believes.

Wael Abd-Almageed, Research Team Lead and Senior Scientist, Information Sciences Institute, University of Southern California and a research team at the University of Southern California Information Sciences Institute developed a study, to distinguish between videos imitation very realistic and original videos that actually happen.

The results of the research found new, more accurate steps to detect deepfake videos:

Find defects

In general, various deepfake algorithms work by using machine learning techniques to analyze images of a target, identifying key elements of facial expressions, such as the nose, the corners of the mouth, the location of the eyes and so on.

The algorithm uses this information to synthesize new images of the person’s face and stitch them together to create a fake video of a target that looks realistic.

The latest method for finding minor flaws in the process of stitching together images in a video deepfake is to view individual frames of the video, either manually or using an algorithm.

If there are a lot of defects in a few frames, it can be said that the video is a deepfake.

However, the creators of deepfake videos began to compress large amounts of images and videos to cover up flaws.

– .