Abstract
Hand-geometry biometric recognition is normally based on the detection of five points that correspond to the fingertips and four points between them (valley points). Specific methods often have to be implemented during the acquisition stage to make the detection of those points easier. This study presents techniques that have been developed to overcome the difficulties and limitations of the current systems. Moreover, a hand-geometry based recognition system that has no constraints during image acquisition is presented. A methodology was developed based on the hand skeleton for the points on the fingertips and for the valley points it was based on the curvature of the hand contour. The principal difficulties were found during the segmentation step, which often fails if the fingers are not spread out. Once the points have been located, the necessary features for authentication were extracted. Classification algorithms were implemented at this stage. Those showing the best results presented a Genuine Acceptance Rate (GAR) of 76% and 8% for the False Acceptance Rate (FAR).

Abstract
In the last few years the research community has witnessed significant progress in biometric technology, due to the availability of a wide variety of databases. However, the available databases that are currently available present significant setbacks in terms of restricted access to data, low-resolution and restrictions imposed on individuals during the acquisition phase. In this paper, two new public databases are described that have been created, with fingerprint and palm print images and their characteristics are compared with other databases available in the research community. The advantages of these databases are the great variety of individual characteristics, they have no restrictions during acquisition and they have manual ground truth annotation. They were presented in two different international competitions and have been used in research by different authors.

Abstract
Omni-directional robots are becoming more and more common in recent robotic applications. They offer improved ease of maneuverability and effectiveness at the expense of increased complexity. Frequent applications include but are not limited to robotic competitions and service robotics. The goal of this work is to find a precise dynamical model in order to predict the robot behavior. Models were found for two real world omni-directional robot configurations and their parameters estimated using a prototype that can have 3 or 4 wheels. Simulations and experimental runs are presented in order to validate the presented work.

Abstract

Abstract

Abstract
Automatic biometric identification based on fingerprints is still one of the most reliable identification method in criminal and forensic applications. A critical step in fingerprint analysis without human intervention is to automatically and reliably extract singular points from the input fingerprint images. These singular points (cores and deltas) not only represent the characteristics of local ridge patterns but also determine the topological structure (i.e., fingerprint type) and largely influence the orientation field. Poincaré Index-based methods are one of the most common for singular points detection. However, these methods usually result in many spurious detections. Therefore, we propose an enhanced version of the method presented by Zhou et al. [13] that introduced a feature called DORIC to improve the detection. Our principal contribution lies in the adoption of a smoothed orientation field and in the formulation of a new algorithm to analyze the DORIC feature. Experimental results show that the proposed algorithm is accurate and robust, giving better results than the best reported results so far, with improvements in the range of 5% to 7%. © 2009 IEEE.