In a world where tens of millions of people use “12345” or “qwerty” as their passwords, it makes sense to use biometric security instead, since each person has several unique biometric signatures that can be used instead of a password. Popular thinking about biometric security often includes fingerprint readers, iris or retina readers, and voice-activated systems. Research published in the International Journal of Biometrics has demonstrated how human laughter can be used in biometrics.
Scientists at the Systems Engineering Department of the University of Lagos in Akoka, Lagos, Nigeria, have found that people can identify other people by the incomparable nature of their laughter. Unlike voice and manner of speech, laughter can hardly be imitated.
The researchers then identified different sound frequencies in a person’s laughter that can be used to create a digital signature, akin to the little things (hashes) generated by fingerprint scanners.
On-site trials show that their prototype recognition algorithm has a 90% accuracy, which compares very favorably with the 65% accuracy of the conventional Gaussian noise model. However, combining the developer’s algorithm with a Gaussian approach can improve overall accuracy by more than 5 percent.
Thus, laughter is a viable biometric function for personal identification that can be embedded in artificial intelligence systems in various applications, the team concludes.