1. AbstractThis article describes a new technique to authenticate handwritten signatures using offline mode. In this technique, every single pixel that belongs to the signature is considered. All edge points/end points are extracted from the signature. These edge points/end points are connected to form a closed polygonal shape. From the polygonal shape, several values ​​can be calculated that can serve as structural characteristics: shape factor, circularity measure, rectangularity measure, minimum rectangle, area and perimeter. These values ​​combine together to create a verification function that can discriminate between genuine and forged signatures.2. IntroductionRecognition of a person can be carried out both on the basis of behavioral and physical characteristics in automated biometric methods. There are many behavioral attributes which can be voice, iris, fingerprint and facial recognition. Due to the increase in fraud resulting from counterfeits, the need to develop secure systems to authorize the right person has increased, and these systems need to be more sensitive to discriminate between genuine and counterfeit people. Among the different identification methods, the common method used in our society to identify a person is through handwritten signature because it is an official/formal way of identifying the person. They are used in government, for attestations, authenticity of documents etc. But with his social acceptance he is demoralized by the falsification of making fake transactions. The need is to minimize forgery threats, research has been conducted and it is still an interesting field for researchers to minimize the acceptance of forged signatures. Automated verification...... half of the document ...... n Methods for offline Signature Verification”, Frontiers in Handwriting Recognition, 2004. IWFHR-9 2004, pages 161–166.[18] Justino, E., Bortolozzi, F. and Sabourin, R. (2005), “A comparison of SVM and HMM classifiers in off-line signature verification”, Pattern Recognition Letters, 6(9):1377–1385.[ 19] ¨Ozg¨und¨uz, E., S¸ent¨urk, T., and Karslıgil, M. (2005), “Off-Line Signature Verification and Recognition by Support Vector Machine,” in European Conference on Computing of the signal.[20] ] Ma, Z., Zeng, - Class-One-Network", LECTURE NOTES IN COMPUTER SCIENCE, 4493:1077[21] Srihari, S., Srinivasan, H., Chen, S. and Beal, M. (2008), "Machine Learning for Signature Verification" , Intelligence (SCI), 90:387–408.