Optimization approaches in DTW based on-line siganature verification

OData support
Supervisor:
Dr. Kővári Bence András
Department of Automation and Applied Informatics

The automatic signature verification helps to programmatically decide about a signature’s genuineness without the presence of experts. In everyday life there are a lot of situations in which this can be very useful. The target is to build an accurate and trusted verification system.

This paper deals with on-line signature verification, within which it uses DTW algorithm to compare signatures. The verification system requires a test database, based on which it can be determined how similar are the genuine signatures. A new signature will be evaluated using this information. The dataset was collected and stored as it is common in on-line signature verification. The evaluation method discussed in this paper uses only those pieces of available information which are also accessible on pictures of a signature.

The evaluation method discussed in this paper reached Average Error Rate (AER) of about 25-30%. The measured minimal value was 15% and the maximal was 57%. The results give countenance to the significance of this investigational area, but they still can be optimized.

Downloads

Please sign in to download the files of this thesis.