Biometric identification by detecting hand veins in infrared images

Abstract

This paper presents a methodology to implement a computational system that can identify people by the geometry of the veins of the dorsal hand from infrared images. The methodology is as follows: First, a prototype is built by a web camera whit infrared LEDs that allow capture infrared images of dorsal hand. Once the images are captured, the image analysis algorithms are applied by digital image processing in order to segment regions of interest. Once segmented images are obtained characteristics and the pattern representing each input image is formed. Since the input pattern for each image is obtained respectively, associative memory model based on cellular automata model responsible for pattern classification is constructed.