Touch-less Fingerprint Recognition UsingSVM and GMM: A Comparative Study (original) (raw)
Related papers
Touch-less Fingerprint Recognition System
Touch-less Fingerprint Recognition System is a viable alternative to conventional touch based fingerprint recognition system. Fingerprint recognition is one of the application of biometrics that is used for identification of a person. In touch based sensing, many sensors were developed for fingerprint recognition purpose in which the user's fingerprint is placed on the sensor. Depending upon the pressure applied on a sensor by the person the input fingerprint from one and same finger can vary which may lead to problems like forgery and hygiene. Touch-less fingerprint technology is developed so that the problems in touch based sensing techniques can be depleted as this system avoids physical contact between a finger and sensor. Touch-less system is different from conventional system because they use digital camera to acquire the fingerprint image whereas in conventional system live acquisition technique is used. In touch-less fingerprint system we consider constraint of the fingerprint images that were acquired with the digital camera such as the low contrast between the ridges and valleys in the fingerprint images, motion blurriness and defocus. Touch-less system can be mainly divided into four major parts they are data acquisition, pre-processing, extraction of features and matching. In feature extraction, minutia from fingerprint images are extracted and in matching process the number of minutia pairings between two fingerprints is matched. This project is coded by using MATLAB software. Fingerprint recognition system has been widely adopted for verification purpose because of their reliability as compared to other biometric application. As fingerprint is believed to be unique for each person fingerprint recognition has found its application in various different fields.
Review of Touch-Less Fingerprint Recognition Technologies
Fingerprint recognition has aided variety of biometric security applications over past few decades. Each person possesses unique fingerprint characteristic in terms of minutia, pore, ridges and patterns. Based on data acquisition methods fingerprint recognition can be touch based or touch less, with later having advantage in terms of better hygiene, safety and stray fingerprints. In Touch-less fingerprint recognition, fingerprints are acquired using a high resolution digital camera or any other optical acquisition system. Such systems find applications in numerous fields such as secure access to laptops, computer systems, cellular phones, banking, ATMs etc. Deceiving the simple appearance various taxing, such as non-uniform light, movement blurriness, defocus and low contrast between ridges and valley etc. To overcome such problems, it is important to focus more on the pre-processing steps. Touch-less fingerprint detection can be classified as two-dimensional and three-dimensional methods depending upon number of cameras used for acquisition of finger-print image. Literature presents different techniques for acquisition and analysis of finger-print images. In this paper, an effort is being made to review some of these techniques and give a brief comparison for the same in terms of accuracy, merits, demerits and their respective solution.
Support Vector Machine Based Fingerprint Identification
This work is released in biometric field and has as goal, development of a full automatic fingerprint identification system based on support vector machine. Promising Results of first experiences pushed us to develop codification and recognition algorithms which are specifically associated to this system. In this context, works were consecrated on algorithm developing of the original image processing, minutiae and singular points localization; Gabor filters coding and testing these algorithms on well known databases which are: FVC2004 databases & FingerCell database. Performance Evaluating has proved that SVM achieved a good recognition rate in comparing with results obtained using a classic neural network RBF.
Fingerprint classification decreases the number of possible matches in automated fingerprint identification systems by categorizing fingerprints into predefined classes. Support vector machines are widely used in pattern classification and have produced high accuracy when performing fingerprint classification. In order to effectively apply Support vector machines to multi-class fingerprint classification systems.It is proposed a novel method in which the fingerprint classification can be done by the classifier used Naïve Bayes and Support vector machines efficiently reduce the search time by restricting the subsequent searching stage to either left hand thumb and right hand thumb databases.
Evaluation of a Fingerprint Recognition Technology for a Biometric Security System
2018
Authentication is a fundamental component of human interaction with computers. Traditional means of authentication, primarily password and personal identification numbers (PINs), have until recently dominated computing, and are likely to remain essential for the years to come. Over the years, passwords are kept simple to avoid them being easily forgotten. This has subjected them to higher vulnerability to much compromise by unauthorized persons. Thus, computers are forced to manage more and more passwords which imply that the likelihood of password being forgotten increases. Hence, biometrics is becoming more convenient and distinctly more precise than traditional methods such as passwords and PINs. Biometrics link the event to a particular individual, requires nothing to remember or carry, provides positive confirmation by verifying individuals are who they claim to be, and is becoming an inexpensive solution. This research work aims to evaluate fingerprint recognition technology f...
Biometrics Fingerprint Sensors: An Introduction
International Journal for Scientific Research and Development, 2015
Every human has distinct physiological as well as behavioral characteristics and to recognize each individual we need a biometric system that provides a wide variety of reliable personal recognition schemes either in form of confirmation or determination. The purpose of such schemes is to make sure that the services are only accessed by a genuine user, and no one else use of it. This includes fingerprint, voice, face, gait, iris, signature, hand geometry etc. Mainly the biometric system consists of four main modules i.e. sensor module, feature extraction module, matching module and decision module. In a biometric system, sensor module is the first module so it is very essential to have accurate acquisition of data over there. In this paper various type of fingerprint sensors are discussed in detail. There are some parameters like pixel density, resolution, SN ratio, motion blur, gray scale etc. of these sensors which are also included. Sensors also encounter some challenges like dur...