Off-line Signature Verification based on Hu's Moment Invariants and Zone Features using Support Vector Machine
Mandeep Kaur Randhawa, Guru Nānak · 2012
An off-line signature verification system that incorporates a novel feature extraction technique is proposed in this paper. Moments of images have been used expansively in image analysis applications. In this paper, the fusion of Hu's moment invariants and zone features extracted from signature images is used as input patterns. Support Vector Machine (SVM) technique is used to verify the system. From the experimental results, the new features proved to be more robust than other related features used in the earlier systems. The proposed system has 4% errors in rejecting skilled forgeries (FRR), 2% errors in accepting genuine signatures (FAR) and 4% errors for identifying the particular person signature as some other signature (FIR). In all, the model exhibited 90% acceptance rate in terms of percent acceptance factor.