Results (
Thai) 1:
[Copy]Copied!
This paper aims to verify offline signatures usingimproved feature analysis and artificial neuralnetwork. Feature analyzer can reduce the largedomain of feature space and extract invariableinformation. We incorporated different features frommulti-dimensional feature analysis perspective. Forverification from extracted features, we used neuralnetwork classifier. Instead of using feed forwardneural network, multiple feed forward neural networksare used which are trained in the form of ensemble.Using such ensemble makes the system more generalthan a regular single neural network based system.Use of resilient back propagation for each neuralnetwork training, provides faster recognition. Usingcross validation techniques, we performed significantamount of testing. Experimental evaluation of thesignature verifier is reported.
Being translated, please wait..
