With the introduction of the internet, most of the people turn to the social media for the news and information. However, recent development in the field of technology has completely changed the way we look at online information. Advanced technology has made it very easy to fake a news which has become one of the most challenging things for almost every social media network. Companies and Fact-checkers are working vigorously to verify the news before publishing however, on such a large scale verification process can be very time consuming and expensive.
There are different deepfake detection software that help in the detection of a fake video by checking and assessing frame to frame difference in the video. However, a new tool has been introduced that made it very easier to detect the fake video. The use of this artificial-intelligence requires less energy and less time because it assesses the whole video at once. The process is simple, this AI does not asses frame to frame difference, instead, it arranges all the frames one by one on top of each other. Once all the frames are stacked, it then determines any difference between the frame that can help in the detection of the fake video. According to the researchers at the USC Information Sciences Institute (USC ISI), this new digital forensics technique is up to 90% accurate and it has been trained by extracting data from 100 different manipulated videos.
The basic trick that this software uses is to determine any soft-biometric signature. This means that any inconsistency in the frame of the video will help in determination is the video is fake or real. This difference can be as small as the difference in eyelid movement or as big as a change in the environment. Experts are hopeful that with the help of this software that has been designed by USC ISI, they will be able to extract leads for the 2020 presidential election. This will extract and minimize all the doubts related to the negativity and election hacking. This software will help in stopping the spread of fake videos and fake news and becoming this fake news part of the newsfeed as well a becoming viral.
Read next: Deepfake — A technological blessing or a curse
There are different deepfake detection software that help in the detection of a fake video by checking and assessing frame to frame difference in the video. However, a new tool has been introduced that made it very easier to detect the fake video. The use of this artificial-intelligence requires less energy and less time because it assesses the whole video at once. The process is simple, this AI does not asses frame to frame difference, instead, it arranges all the frames one by one on top of each other. Once all the frames are stacked, it then determines any difference between the frame that can help in the detection of the fake video. According to the researchers at the USC Information Sciences Institute (USC ISI), this new digital forensics technique is up to 90% accurate and it has been trained by extracting data from 100 different manipulated videos.
The basic trick that this software uses is to determine any soft-biometric signature. This means that any inconsistency in the frame of the video will help in determination is the video is fake or real. This difference can be as small as the difference in eyelid movement or as big as a change in the environment. Experts are hopeful that with the help of this software that has been designed by USC ISI, they will be able to extract leads for the 2020 presidential election. This will extract and minimize all the doubts related to the negativity and election hacking. This software will help in stopping the spread of fake videos and fake news and becoming this fake news part of the newsfeed as well a becoming viral.
Read next: Deepfake — A technological blessing or a curse
