Some observations on practical exploitation of automatic signature verification technologies

Despite the long history of research in automatic signature verification, attempts to realise systems robust enough for routine practical use have not yet met with the degree of success originally expected. Although the principal reason for this generally cited by potential users is that performance levels attainable are insufficiently high, it is clear that there are at least three interrelated issues which must be addressed if practical success is to be achieved. These are: optimal system design; optimal operational integration of the system; appropriate system evaluation. This paper considers each of these issues and presents illustrative examples of how these important problems may be addressed, with a view to creating a system design which maximises the potential for practical use