Researchers develop a new technique that will keep your online photos safe from facial recognition algorithms
Facial Recognition tools are one of the foremost tools used by law enforcement authorities to keep a check on protestors and activists. Such tools can even be used by stalkers and hackers to identify you from that one image you posted online. Even social media networks like Facebook and Instagram can automatically tag your photos using AI and Google uses its own facial recognition tech for various purposes.
In short, facial recognition tools are a serious threat to your privacy. In fact, many states in the United States are banning any use of facial recognition tech. But that doesn’t stop either the usage of facial recognition tech or its evolution. However, there is a way out thanks to a new technique developed by Professor Mohan Kankanhalli, Dean of the School of Computing at the National University of Singapore, and his research team.
The research, which has been ongoing for more than six months, is targeted at countering the facial-recognition algorithms of big tech firms such as Facebook and Google. The NUS technique stops such artificial intelligence (AI) software from recognizing specific facial attributes, such as gender and race, by introducing subtle visual distortions in the image. In fact, the changes made to your image using this technique are so minor that you wouldn’t even notice it with your eyes but the changes can easily make it hard for facial recognition tech to recognize your images.
“It’s too late to stop people from posting photos on social media. However, the reliance on AI is something we can target,” says Dr.Kankanhalli. He says that the NUS Department of Computer Science technique safeguards sensitive information in photos by making subtle changes that are almost imperceptible to humans but render selected features undetectable by known algorithms.
Dr.Kankanhalli and his team have open-sourced the unnamed technique developed by him and his team. The NUS team took six months of research to develop this novel technique, and their achievements were published in the Proceedings of the 27th ACM International Conference on Multimedia. The team also plans to extend this technology to videos, which is another prominent type of media frequently shared on social media platforms.
The team also plans to extend this technology to videos, which is another prominent type of media frequently shared on social media platforms. The source code of the algorithm is available on the NUS website for developers to incorporate into their apps.