Please use this identifier to cite or link to this item: http://ea.donntu.edu.ua:8080/jspui/handle/123456789/22918
Full metadata record
DC FieldValueLanguage
dc.contributor.authorФедоров, Е.Е.-
dc.contributor.authorFedorov, Ye.Ye.-
dc.contributor.authorФедоров, Є.Є.-
dc.date.accessioned2013-09-24T11:25:33Z-
dc.date.available2013-09-24T11:25:33Z-
dc.date.issued2013-
dc.identifier.citationНаукові праці Донецького національного технічного університету. Серія: Обчислювальна техніка та автоматизація. Випуск 2 (25). - Донецьк, ДонНТУ, 2013. С - 270-278en_US
dc.identifier.issn2075-4272-
dc.identifier.otherУДК 004-
dc.identifier.urihttp://ea.donntu.edu.ua/handle/123456789/22918-
dc.descriptionNow the system, intended for biometric person identification applied in criminalistics and various security systems, is more important. In the article a method of biometric identification of a person based on digital processing of a signal, recognition of speech and visions, neural networks, the fuzzy logic and genetic algorithm has been offered. The method of biometric identification provides: formalisation of signs of speech; formalisation of image features; creation of adaptive neuro-fuzzy inference system of biometric identification; construction of genetic algorithm. As speech features the frequencies of first three formant vowel sounds have been chosen. As a person’s face features 3 points for the left eyebrow (a head, a break point, an eyebrow tail) have been chosen; 3 points for the right eyebrow (a head, a break point, an eyebrow tail); 4 points for the left eye (left, right, top, bottom); 4 points for the right eye (left, right, top, bottom); 4 points for a nose (nose bridge, the nose basis, the left and right wing of a nose); 4 points for a mouth (left, right, top, bottom); 1 point for a chin have been chosen. Effective variants of operators of reproduction (a combination of casual and linearly ordered selection with annealing imitation), crossingover (for the choice of parents the combination of outbreeding and inbreeding with imitation annealing is used), mutation(a combination of homogeneous mutation with imitation annealing) and reduction (a combination of the equiprobable and selective scheme with annealing imitation) have been offered. This encreases the efficiency of training of the systems of biometric identification. This method has been numerically investigated on standard basis TIMIT (for identification of a person by speech) and ORL (for identification a person’s face) and in its probability of recognition and speed of training this method is comparable to the best methods. The developed hybrid intellectual computer system of biometric identification of the person possesses the following advantages: possibility of use of the aprioristic information (knowledge of experts); representation of knowledge in the form of the rules easily accessible to understanding by the person; possibility of fast training and adaptation; parallel processing of the information which raises computing capacity; there are no difficulties with definition by system structure; high probability of identification.en_US
dc.description.abstractДля создания интеллектуальной компьютерной системы биометрической идентификации человека в статье был предложен метод на основе формантных признаков речи, антропометрических признаков лица человека, искусственной нейронной сети, нечеткой логики и генетического алгоритма. Были предложены архитектура нечеткой искусственной нейронной сети и эффективные варианты операторов генетического алгоритма (репродукции, кроссинговера, мутации и редукции) на основе имитации отжига, которые позволяют учитывать этапы генетического алгоритма. Разработанный метод был исследован на стандартных базах TIMIT и ORL.en_US
dc.publisherДонецький національний технічний університетen_US
dc.subjectформантні ознаки мовленняen_US
dc.subjectантропометричні ознаки особиen_US
dc.subjectнейронна мережаen_US
dc.subjectнечітка логікаen_US
dc.subjectгенетичний алгоритмen_US
dc.subjectformant speech featuresen_US
dc.subjectanthropometrical features of a personen_US
dc.subjectneural networken_US
dc.subjectfuzzy logicen_US
dc.subjectgenetic algorithmen_US
dc.subjectформантные признаки речиen_US
dc.subjectантропометрические признаки лицаen_US
dc.subjectнейронная сетьen_US
dc.subjectнечеткая логикаen_US
dc.subjectгенетический алгоритмen_US
dc.titleРАЗРАБОТКА МЕТОДА БИОМЕТРИЧЕСКОЙ ИДЕНТИФИКАЦИИ ЧЕЛОВЕКАen_US
dc.title.alternativeDeveloping a Method of Biometric Identification of a Personen_US
dc.title.alternativeРозробка методу біометричної ідентифікації людиниen_US
dc.typeArticleen_US
Appears in Collections:Випуск 2 (25)'2013

Files in This Item:
File Description SizeFormat 
федюн.pdf434,39 kBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.