I've seen accuracy [(2-FAR-FRR)*100/2] reaching to about 80% to 85%. Could anyone please share with me the greatest accuracy you know in off-line HSV and also specify the reference from which the value was obtained?

Also do any of these methods address the individual variability of signers? Eg., an expert signer would have a greater self similarity in his/her signature, while a novice signer would have a low self similarity in signing his/her own genuine signature. Can any of those highly accurate HSV systems retain their accuracy in both these cases? If so, how? Are there any other limitation to such systems?

I do believe the contribution to these questions would benefit many in this field of study.

More Ebenezer R.H.P. Isaac's questions See All
Similar questions and discussions