Design and implementation of biometric recognition using off-line signature analysis

Analysis of signature is a widely used and developed area of research for personal verification. A Software is introduced to solve problems related with verification of individuals by using off-line signature verification which is a branch of biometric verification in the first part of study. The presented signature verification system includes image processing techniques and other mathematical tools in its structure. The proposed system is composed of five sub-systems; data entry, preprocessing, feature extraction, comparison and performance evaluation. The developed system tries to determine the originality of the signature mathematically. The technique is based on Neural Network (NN) approach trained with Particle Swarm Optimization (PSO) algorithm. To test the performance of the proposed algorithm, three types of forgeries; random, unskilled and skilled are examined and the experimental results are presented with verification accuracy and statistical figures.Design and construction of a questioned document investigation device is also included in the second part of study. Varieties of document examinations are facilitated by special lightening features and lens systems. This questioned document device can perhaps be one example of an optomechatronics system which includes cameras, different kind of light sources (ultraviolet and infrared) and lenses. Using modern spectral imaging technology it has been possible to detect very small differences available on inks and papers. Image Processing Software for Questioned Document (IMPQD) is also used to control of both the designed device and perform signature verification using the image of document and an available signature database.