Open AccessCosine similarity for analysis and verification of static signatures
Author(s) -
Pirlo Giuseppe,
Impedovo Donato
Publication year - 2013
Publication title -
iet biometrics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.434
H-Index - 28
eISSN - 2047-4946
pISSN - 2047-4938
DOI - 10.1049/iet-bmt.2013.0012
Subject(s) - computer science , signature (topology) , cosine similarity , data mining , stability (learning theory) , similarity (geometry) , matching (statistics) , process (computing) , feature (linguistics) , pattern recognition (psychology) , digital signature , artificial intelligence , algorithm , image (mathematics) , machine learning , mathematics , programming language , hash function , linguistics , statistics , philosophy , geometry
The stability of handwritten signatures is a crucial characteristic for both investigating the nature of the signature apposition process and improving systems for automatic signature verification. In this study, a new technique for the analysis of stability in static signature images is discussed. The technique adopts a feature‐based strategy to derive regional information from a static signature image and uses cosine similarity to estimate the degree of regional stability among genuine signatures, according to a multiple matching strategy. The experimental test carried out using signatures in the Grupo de Procesado Digital de Senales (GPDS) database has demonstrated the validity of this novel approach in obtaining stability information and deriving significant signer‐independent and signer‐dependent properties of the signing process, useful for verification aims.