KeyLemon Blog

In the field of facial recognition

The security of private information is on the minds of everyone who uses a computer. Facial recognition has been introduced to allow the use of facial features instead of passwords. In this article, we are going to discuss the various aspects of this technology and its evolution.

The algorithm up to now

The first problem people mention with respect to facial recognition is identity theft. The possibility that someone could use a photo or a digital device to pass himself or herself off as you.

In the case of KeyLemon, this question has often been raised. Our clients are afraid to find their computer violated by a malicious third party who simply accessed their photo. This problem is more and more frequent and practicable. Since services like Facebook and Flickr have become so widespread, it has become relatively simple to find a photo of one’s self online.

Several solutions for photo attacks have been thought up. The disadvantage is that the user has to initiate an action to prove that he or she is real. Voice recognition coupled with facial recognition could offer optimal assurance. The problem is that the interaction between voice and face involves an increase in identification time. This we are committed to reduce. What’s more, these types of solutions require more computer resources.

What’s new with current algorithm?

The facial recognition algorithm is continually improving. Henceforth, it is possible to detect if you are dealing with a real person or an imposter. We have seen results that demonstrate that the majority of attacks by photo or photo phone are rejected through recognition.

In the following video you can see this improved detection permitting the exclusion of photo attacks.

With KeyLemon we have the possibility of integrating this new algorithm. It will be available with the next release.

What is KeyLemon’s position?

KeyLemon wants first and foremost to provide users with a simple and practical way of logging in to their computer. We have always defined our position in terms of improving the interaction between man and machine.

Conclusion

The algorithm development will not stop at this stage. Performance will continue to be improved alongside improvements to user webcams.