Barcode Generation For Face Images

Abstract

     The paper addresses the problem of linear barcode generation for face images. The history of the problem and possible approaches to its solution in mobile oriented systems are discussed.
     Two methods are presented: the first one is based on intensity histograms, and second one is based on intensity gradients, calculated over images using their original features. Then these features are averaged over limited number of intervals, quantized in the range of decimal digits from 0 tо 9 and converted into standard barcode. Structure of barcode generation system is proposed, and description of its blocks is presented.
     The methods have been tested by using «Face94», «Face Sketch FERET Database» databases, as well as a database composed of different age faces. The tests have demonstrated invariance of barcode in respect to changes in local sizes of face images, in tilt in the XY plane, to changes in the view, mirror rotation about vertical axis, as well as changes in facial expressions and age-related face changes.
     Therefore, the presented methods constitute novel solutions to practical applications in real conditions of dynamical parameters changes in face images. Moreover, both methods require neither big computational resources, nor application of special software packages for image processing, allowing generation of linear barcode in real time systems. A generated standard barcode contains information about a person’s face and can be used for indexing, identification, recognition, and searching for people.