Newly developed software detects 'deepfake' images and videos
- 3 years ago
잡았다 '합성 사진'…위•변조 잡아내는 소프트웨어 국내 최초 개발
With the use of modern technology, doctored videos and images can be made to look like they are real.
These are called 'deepfakes' and are becoming a problem.
But now, there is a new type of software that can detect whether a picture or video has been edited.
The aim is to prevent certain types of digital sex crimes such people having their faces being superimposed into inappropriate, often pornographic material.
Jang Tae-hyun explains.
A video of dinosaurs roaming the streets.
Is this real or fake?
'KAICATCH', a new type of software, developed by researchers from the Korea Advanced Institute of Science and Technology, can determine exactly that.
The software was recently put into practical use for the first time in South Korea, and only the second time globally.
It can find every minor change by analyzing the pixels... like every copy, paste or delete.
The green part indicates that this section has been edited.
"The previous software could not detect random images. Out of 100 pictures, the accuracy was 5 to 10 percent. However, this new software has 70 to 80 percent accuracy."
'Deepfake' is a technique that uses artificial intelligence to edit faces within pictures or videos to get as close to reality as possible.
This can be a problem if used inappropriately.
For example, some people might superimpose a celebrity's face into a pornographic video... an issue which has become prevalent with female K-pop stars in particular falling victim to these types of crimes.
To prevent this from happening, the research team collected around 300-thousand pictures over a five-year period to help improve its detecting skills.
"We used artificial intelligence to develop the software. Previously, we could only detect one change with one type of technology. But now, we can detect up to 5 kinds or 20 algorithms at once and even find out which part has been edited."
The research team is planning to commercialize the program by the end of this year... and aims to prevent crimes related to deepfakes.
Jang Tae-hyun, Arirang News.
With the use of modern technology, doctored videos and images can be made to look like they are real.
These are called 'deepfakes' and are becoming a problem.
But now, there is a new type of software that can detect whether a picture or video has been edited.
The aim is to prevent certain types of digital sex crimes such people having their faces being superimposed into inappropriate, often pornographic material.
Jang Tae-hyun explains.
A video of dinosaurs roaming the streets.
Is this real or fake?
'KAICATCH', a new type of software, developed by researchers from the Korea Advanced Institute of Science and Technology, can determine exactly that.
The software was recently put into practical use for the first time in South Korea, and only the second time globally.
It can find every minor change by analyzing the pixels... like every copy, paste or delete.
The green part indicates that this section has been edited.
"The previous software could not detect random images. Out of 100 pictures, the accuracy was 5 to 10 percent. However, this new software has 70 to 80 percent accuracy."
'Deepfake' is a technique that uses artificial intelligence to edit faces within pictures or videos to get as close to reality as possible.
This can be a problem if used inappropriately.
For example, some people might superimpose a celebrity's face into a pornographic video... an issue which has become prevalent with female K-pop stars in particular falling victim to these types of crimes.
To prevent this from happening, the research team collected around 300-thousand pictures over a five-year period to help improve its detecting skills.
"We used artificial intelligence to develop the software. Previously, we could only detect one change with one type of technology. But now, we can detect up to 5 kinds or 20 algorithms at once and even find out which part has been edited."
The research team is planning to commercialize the program by the end of this year... and aims to prevent crimes related to deepfakes.
Jang Tae-hyun, Arirang News.