Robust on-line signature verification by new segmentation method and fusion model

This paper proposes a robust on-line signature verification by a new segmentation method and fusion model. The segmentation method solves the problem that the variation between reference signature and input signature causes an error in the location or the number of segments. In addition, the fusion model discriminates genuineness by calculating each feature vector's fuzzy membership degree yielded by the proposed segment method. Experimental results show lower FRR(False Reject Rate) for genuine signature when FAR(False Accept Rate) is fixed than the dynamic programming(DP) method.