On the relevance of age in handwritten biometric recognition

One of the main drawbaks of biometric recognition is its unstability along time. One of the main reasons for the loss of performance of biometric systems is the aging of its enrolled users. Features -and models built from them- that were up-to-date at the time of their acquisition may eventually become outdated, leading to a failure in recognition tasks. In this paper, we analyze the effect of aging in online handwriting as well as its relevance on biometric error rates. Experimets have been conducted on our BIOSECURID database, which contains data donated by 400 individuals. According to our preliminary experiments, the average handwriting speed steeply decreases for writers who are 60 and above. Regarding the relation between verification errors and age, results suggest that false rejections are not age-dependent while false acceptances seem to be related to the age of the writer.