An Improved Method for Segmentation of Fingerprint Images

Abstract

In this work, an improved method for the pixel-wise segmentation technique is presented. Three features are extracted from each pixel and a linear classifier associates the pixel with the background or the foreground. Since this technique requires a trade-off between the accuracy of the classification and the computational effort, we propose a modified editing-condensing technique to select a reduced and representative reference set from the original training set. Also, because this is a linearly nonseparable classification problem, we propose the fuzzy perceptron learning method to obtain an optimal and robust lineal classifier. Experiments have shown that using the proposed method a reduced number of epochs and classification errors were obtained