Handwritten signature verification using hidden Markov models

[摘要] ENGLISH ABSTRACT:Handwritten signatures are provided extensively to verify identity for all types of transactionsand documents. However, they are very rarely actually verified. This is because ofthe high cost of training and employing enough human operators (who are still fallible) tocope with the demand. They are a very well known, yet under-utilised biometric currentlyperforming far below their potential. We present an on-line/dynamic handwritten signatureverification system based on Hidden Markov Models, that far out performs humanoperators in both accuracy and speed. It uses only the local signature features-sampledfrom an electronic writing tablet-after some novel preprocessing steps, and is a fullyautomated system in that there are no parameters that need to be manually fine-tunedfor different users. Novel verifiers are investigated which attain best equal error rates ofbetween 2% and 5% for different types of high quality deliberate forgeries, and take afraction of a second to accept or reject an identity claim on a 700 MHz computer.

[发布机构] Stellenbosch University