The basic aim of a biometric identification system is to discriminate automatically between subjects in reliable and dependable way, according specific-target application. Multimodal systems fuse two or more physical behavioral traits provide optimal False Acceptance Rate (FAR) Rejection (FRR), thus improving accuracy dependability. In this paper, an innovative multimodal based on iris fingerprint proposed. paper state-of-the-art advancement multibiometrics, offering perspective features...

Medical imaging involves important computational geometric problems, such as image segmentation and analysis, shape approximation, three-dimensional (3D) modeling, registration of volumetric data. In the last few years, Conformal Geometric Algebra (CGA), based on five-dimensional (5D) Clifford Algebra, is emerging a new paradigm that offers simple universal operators for representation solution complex problems. However, widespread use CGA has been so far hindered by its high dimensionality...

Automatic road sign recognition systems are aimed at detection and of one or more signs from real-world color images. In this research, detected extracted real world scenes on the basis their shape features. A dynamic region growing technique is adopted to enhance segmentation results obtained in HSV space. The based a threshold that reduces effect hue instability due external brightness variation. Classification then performed candidate regions using multilayer perceptron neural networks....

Driver's drowsiness is stated as an important cause of road and highway accidents. Therefore, the development a system for monitoring driver's level fatigue desirable in order to prevent The paper presents design implementation able find evidence driver ordinary motor vehicle, car system, made up installed infrared video camera connected Celoxica RC203E FPGA based board, perform real time stream processing. exploits "bright pupil" phenomenon produced by retina, that reflects 90% incident...

Fingerprint classification and matching are two key issues in automatic fingerprint recognition. Generally, recognition is based on a set of relevant local characteristics, such as ridge ending bifurcation (minutiae). global features, core delta singularity points. Unfortunately, points not always present image: the acquisition process ideal, so that broken, or belongs to arch class. In above cases, pseudo-singularity-points will be detected extracted make possible matching. As result,...

The development and the diffusion of distributed systems, directly connected to recent communication technologies, move people towards era mobile ubiquitous systems. Distributed systems make merchant-customer relationships closer more flexible, using reliable e-commerce technologies. These environments need many access points, for creation management secure identities recognition users. Traditionally, these points can be made possible by a software system with main central server. This work...

In this paper a novel embedded fingerprints authentication system based on core and delta singularity points detection is proposed. Typical fingerprint recognition systems use for classification tasks. On the other hand, available optical photoelectric sensors give high quality images with well defined points, if they are present. proposed system, matching position, orientation, relative distance detection. As result, involves comparison between few features leading to very fast rates...

Computer graphics applications require efficient tools to model geometric objects. The traditional approach based on compute-intensive matrix calculations is error-prone due a lack of integration between reasoning and matrix-based algorithms. Clifford algebra offers solution these issues since it permits specification geometry at coordinate-free level. best way exploit the symbolic computing power (Clifford) supporting its data types operators directly in hardware. This paper outlines...

Computer graphics applications require efficient tools to model geometric objects. The traditional approach based on compute-intensive matrix calculations is error-prone due a lack of integration between reasoning and matrix-based algorithms. Clifford algebra offers solution these issues since it permits specification geometry at coordinate-free level. best way exploit the symbolic computing power (Clifford) supporting its data types operators directly in hardware. This paper outlines...

Geometric or Clifford algebra (CA) is a powerful mathematical tool that offers natural and intuitive way to model geometric facts in number of research fields, such as robotics, machine vision, computer graphics. Operating higher dimensional spaces, its practical use hindered, however, by significant computational cost, only partially addressed dedicated software libraries hardware/software codesigns. For low-dimensional algebras, several hardware accelerators coprocessing architectures have...

Fingerprint classification is a key issue in automatic fingerprint identification systems. One of the main goals to reduce item search time within database without affecting accuracy rate. In this paper, novel technique, based on topological information, for efficient described. The proposed system composed two independent modules: former module, Fuzzy C-Means, extracts best set training images, latter C-Means and Naive Bayes classifier, assigns class each processed using only directional...

The present work deals with modern computing systems security issues, focusing on biometric based asymmetric keys generation process. Conventional PKI are private/public generated through RSA or similar algorithms. solution embeds information the In addition corresponding private key depends physical behavioural features and it can be when is needed. Starting from fingerprint acquisition, identifier extracted, cyphered, stored in tamper-resistant smart card to overcome problems of...

User identification is a very interesting and complex task. Invasive biometrics based on traits uniqueness immutability over time. In forensic field, fingerprints have always been considered an essential element for personal recognition. The traditional issue focused full fingerprint images matching. this paper advanced technique recognition partial proposed. This system local analysis micro-features, endpoints bifurcations, extraction. proposed approach starts from minutiae extraction image...

The efficiency and the accuracy of a digital feedforward neural networks must be optimized to obtain both high classification rate minimum area on chip. In this paper an efficient MLP implementation. key features hardware implementation are virtual neuron based architecture use sinusoidal activation function for hidden layer. effectiveness proposed solutions has been evaluated developing different FPGA prototypes energy physics domain automatic road sign recognition domain. decreases...