Neural network based footprint identification without feature extraction

Abstract

In recent years, identification systems with using biometric features are receiving considerable attention. Iris, palmprint, fingerprint and footprint are shown as examples. This paper focused on footprint identification without features extraction. CASIA Database, Dataset-D used for identification database. Dataset-D contain footprint images taken from foot pressure measurement plate. Firtsly, each RGB image converted gray scale and resized the fifth and resized 30×15 matrix. In the end, each 30×15 matrix is converted to 1×450 input array, and simulated by MLP, SVM and Naive-Bayes classifiers. The best result without features extraction achived by MLP classifier.