Unlike other types of biometrics, keystroke dynamics can be captured on a standard computer keyboard and requires no additional hardware. This makes it an attractive authentication method as it can be used to monitor and detect anomalies for real-time authentication of a subject. Moreover, it is highly accurate in detecting changes to the normal typing pattern of a subject.

The speed at which a subject presses keys (flight time) and the duration of holding each key down (dwell time) vary from person to person due to their different typing patterns. This enables software to identify and recognize the individual characteristics of each subject’s typing style, resulting in higher accuracy than traditional biometric methods such as fingerprint or face recognition.

Keystroke dynamics is also a non-intrusive, user-friendly technology that does not require users to perform complex tasks or gestures to use. Instead, it can be implemented in an existing application with no extra effort required by the user and thereby increases user acceptance of the technology. It can also help to prevent fraud or theft by continuously monitoring the typing behavior of a subject and alerting if an anomaly is detected.

Compared to other technologies that are based on physical or behavioral biometrics, the advantages of using dynamic keystroke dynamics include lower cost and simpler integration. Moreover, there are no special sensors needed to capture the subject’s typing patterns as they do not need to be transmitted to a remote server for analysis. Instead, the software-based solution captures the typing data directly from a keyboard and is available to the user immediately without requiring additional hardware such as a smart phone or tablet.

A subject’s typing rhythm varies considerably from day to day, and even within the same day, due to fatigue, switching between computers / keyboards, mood, influences of alcohol and medications and other factors. This sensitivity to changes in the subject’s typing rhythm means that it is difficult to achieve very high FAR and FRR values for behavioral biometrics, such as keystroke dynamics, unless a very large set of samples is collected.

To overcome this, the invention uses a dynamic keystroke prediction engine which compares the selected symbol against a second set of prestored symbol sequences to determine if a match is achieved. When a match is determined in step 36, the microprocessor 4 controls the keyboard 3 to display an alternative keyboard (containing less than all of the original symbols) of the matched dynamic category. The alternative keyboard is preferably embellished to distinguish it from the other non-embellished symbols displayed by the keyboard 3.

For example, in FIG. 10, a user enters the word “anti-climactic.” The prediction engine 74 predicts that the next grammatical character to be entered will most likely be a hyphen, and thus the dynamic key 74 is adjusted to display and represent a hyphen after the entry of ‘anti’.

The alternative dynamic keystroke keyboard of the matched dynamic category can be further altered to show other features that are not displayed by the original symbols displayed on the keyboard, such as icons for certain cellular networks or international calling codes. The icon prediction can further enhance the security of the solution by requiring that the user confirm that they are attempting to authenticate using a valid icon.