Research On Identity Recognition Based On Keystroke Biometrics

With the rapid development of information technology, information security is becoming more and more important, personal authentication technology is receiving people's attention. Because of their inherent defects, the traditional personal authentication cannot meet the fast increasing need of security, biometrics authentication becoming necessary. As an important member of biometrics, keystroke biometrics has attracted extensive research interests in resent years.Traditional keystroke identity recognition system is based on single keystroke trait: keystroke time. But keystroke pressure is also a very important feature of keyboard input, which can perfectly reflect the usr's input habit.This thesis has done a wide range of studies in related field of keystroke recognition. Main contributions are as follows:Firstly, a pressure data collecting system is designed. This thesis designed a pressure sensing keyboard to measure finger-applied pressure during user's input. This device can measure the pressure accurately with a high speed,and transmit the pressure values to computer through COM port. The software on computer save the data to be used in recognition system.Secondly, this thesis researched the time-based keystroke recognition and pressure-based keystroke recognition. A data set has been created using the data collecting system. And some experiments have been done on this data set. Using single feature of time or pressure obtained a experimental result of good recognition rate.Finally, a fusion method to fuse both the time and pressure features has been proposed. The recognition rate has been improved after the fusion.