Journal Homepage: www.ijcst.org

Abstract – Signatures are one of the most important and commonly used tools for human identification. This paper proposes an offline signature verification method based on texture analysis of the image. A sample of signatures is used to represent a particular person. For each known writer sample of fifteen genuine signatures are taken. Forged signatures are also used to test the efficiency of the system. For each signature gray level run length matrix features are extracted and the inter-class distances and intra class distances have been calculated. For each test signature the intra-class threshold is compared to the inter-class threshold for the claimed signature to be verified using Euclidean distance model. Results showed that signature texture feature can be reasonably used for personal verification. Texture based feature extraction technique consistently outperformed the traditional grid based feature extraction technique. Accuracy of 85 % was achieved with the Euclidean distance classifier with FAR and FRR as low 13.33 % and 16.4%