Person authentication with RDTW based on handwritten PIN and signature with a novel biometric smart pen device

This paper presents writer authentication using features of handwritten single characters, PIN words and signatures. The kinematics and dynamics of handwriting movement were recorded with a novel ballpoint pen equipped with a diversity of sensors mounted inside the pen. The time series provided by five different sensor channels including refill and finger grip pressures, acceleration and tilt data was analyzed by using a DTW algorithm. To speed up computation a ldquoreduced dynamic time warping RDTWrdquo technique was applied which is based on the sum of sensor channels and a two steps down-sampling of the multi-dimensional time series. Preliminary results have shown that processing time and memory size could significantly be reduced without deterioration of authentication performance using single characters, signatures and PIN words. Excellent accuracy in recognition was achieved which is mainly due to RDTW technique and the high quality of data sampled by a novel pen device.