Washington: Indian researchers have developed a new keystroke algorithm that can use unique human typing patterns to make online authentication processes more secure, reliable and cheap.
The new method developed by researchers at the Department of Computer Science and Engineering, Jeppiaar Engineering College, Chennai, hopes to alleviate some of the common issues for internet users including loss of password, growing prowess of hackers, and easy access to methods such as phishing and usage of bots.
Like fingerprint scans, retina scans and facial recognition, keystroke dynamics are a biometric - they measure a unique human characteristic.
"As the typing pattern varies from person to person, this can be used as a suitable method for the authentication process more effective than others," researchers J Visumathia and P Jesu Jayarin wrote in the Journal of Applied Security Research.
"The information needed for the process is using the various software systems already present in the computer, leading to a decrease in costs," researchers said.
The new keystroke template algorithm combines measures from existing models to increase precision. To test their algorithm, the researchers built a programme that users could log into using passwords of varying length.
While entering their credentials, keystroke dynamics were recorded.
Results indicate that their algorithm was successful in decreasing login errors and making improper authentication very unlikely, thus advancing keystroke dynamics analysis as a viable e-security measure.
This method is especially appealing for its relative ease of implementation, as the information needed to evaluate human typing patterns is already present in computers, researchers said.
The researchers call for additional testing before the new algorithm can be used as a security measure.
"We concluded from the results presented that keystroke dynamics analysis holds big potential as an authentication method, but the methods used in the process have to be improved before it can be used as an independent security measure," researchers said.