Made available in DSpace on 2016-08-29T15:33:13Z (GMT). No. of bitstreams: 1 tese_4090_.pdf: 1568197 bytes, checksum: 71e4fb308c2516a5a0a305e67f32990f (MD5) Previous issue date: 2010-08-27 / Durante todo o trabalho, o sistema de reconhecimento de fala contínua de grande vocabulário Julius é utilizado em conjunto com o Hidden Markov Model Toolkit(HTK). O sistema Julius tem suas principais características descritas, tendo inclusive sido modificado. Inicialmente, a teoria de reconhecimento de sinais de fala é demonstrada. Experimentos são feitos com adaptação de modelos ocultos de Marvov e com a técnica de validação cruzada K-Fold. Resultados de reconhecimento de fala após adaptação acústica à um locutor específico (e da criação de modelos de linguagem específicos para um cenário de demonstração do sistema) demonstraram 86.39% de taxa de acerto de sentença para os modelos acústicos holandeses. Os mesmos dados demonstram 94.44% de taxa de acerto semântico de sentença.

Magistrinio darbo tikslas buvo nustatyti optimalias parametrų reikšmes automatizuotam garsyno anotavimui, taikant paslėptų Markovo modelių metodiką. Tyrime buvo nagrinėjami 25 skirtingų kalbėtojų įrašai. Kiekvienam kalbėtojui buvo naudojama po 60min įrašų apmokymams ir vienas 2min įrašas buvo naudojamas testavimui. Buvo atliekami keturių tipų eksperimentai pavadinti: PMM modelių apjungimas, PMM tikslinimų skaičius, visų Gauso mišinių įterpimas iškarto, Gauso mišinių įterpimas mišinių skaičių didinat po vieną. Kad palengvinti eksperimentų atlikimą ir rezultatų nagrinėjimą buvo sukurtos dvi programos: garsynų sulyginimo programa ir rezultatų vizualizavimo programa. Garsynų sulyginimo programa sulygino eksperto anotuotą garsyną su automatizuotos anotavimo sistemos anotuotu garsynu. Pagal gaunamus sulyginimo rezultatus buvo galima spręsti apie parinktų parametrų reikšmių gerumą. Gauti sulyginimo rezultatai buvo atvaizduojami rezultatų atvaizdavimo programos pagalba. / In my master degree work I am trying to optimize values of parameters for automatic sound annotation, by using hidden Markov models method. In this research were analyzed 25 different speakers. It was used 60min of speak records for learning and 1 record of 2min speak for testing of all speakers. In this research were used four types of experiments which were named: combination of HMM models, HMM correction number, insertion of all Gaussian mixtures at once, insertion of Gaussian mixtures by increasing mixtures one by one. There was created two programs for facilitating experiment execution: corpus comparison program, result visualization program. Corpus comparison program compares two corpuses expert annotated corpus and automatic annotation system’s annotated corpus. From these results we find out how good are the parameter values. The comparison results can be visualized with the visualization program.

The master thesis is divided into two parts theoretical and practical. The theoretical part is focused on methods of analysis and detection of speech signals. In the practical part the system for isolated word recognition was created in Matlab. The system is speaker independent separately for men and women. Also two speech databases were created for further use in the aircraft cockpit. Tests and evaluations were performed even with added noise.

Magistro baigiamąjį darbą sudaro 51 puslapis. Jame yra 5 lentelės, 36 paveikslėliai, 49 literatūros šaltiniai. Magistro darbe analizuojami vegetacijos pradžios ir žydėjimo pradžios datų kitimai 1991, 2000, 2006, 2007 ir 2008 metais, atliekant kiekvienų metų kritulių ir temperatūros dinamikos analizę. Darbo tikslas: Įvertinti Kamšos draustinyje plačialapių – lapuočių miškų žolinių augalų vegetacijos ir žydėjimo kaitą skirtingomis klimatinėmis sąlygomis. Darbo objektas: Kamšos botaniniame – zoologiniame draustinyje augantys augalai: Lamiastrum galeobdolon, Aegopodium podagraria, Pulmonaria obscura, Polygonatum multiflorum, Stachys sylvatica, Ficaria verna, Mercurialis perennis, Stellaria holostea, Carex sylvatica, Phyteuma spicatum. Darbo metodai: Tyrimui panaudoti 1991, 2000, 2006, 2007ir 2008 metų meteorologiniai ir fenologiniai rodikliai Rudenį parenkama dešimt 1x1 m² dydžio barelių, kurie išdėstomi taip, kad kuo įvairesnės augalų rūšys užimtų nustatytą plotelį. Kiekvienais metais visą vegetacijos periodą nuo pavasario iki rudens kartą per savaite registruojama bareliuose augančių rūšių aukščiai, projekciniai padengimai ir fenologinės fazės. Registruojant nurodoma mėnuo ir diena. Analizuojami kritulių ir temperatūros duomenys. Duomenų grafiniam vaizdavimui panaudotas Microsoft „Office Xp“ programinis paketas. Darbo rezultatai: Tyrimo rezultatai parodo, kad kiekvienas augalas skirtingai reaguoja į temperatūros ir kritulių kitimus pagal savo poreikį. Ankščiausiai augalai... [toliau žr. visą tekstą] / Master work consists of 51 pages. It contains 5 tables, 36 illustrations, 49 references. Master's work analyzes vegetation and flowering dates of the beginning of variation 1991, 2000, 2006, 2007 and 2008, during each year of precipitation and temperature dynamics in the analysis. The purpose of work: To evaluate the Kamsa rezerve the broad – hardwood forest herbaceous plant vegetation and flowering change under different climatic conditions. The subject of work: The Kamsa reserve with herbaceous plants: Lamiastrum galeobdolon, Aegopodium podagraria, Pulmonar obscura, Polygonatum multiflorum, Stachy sylvatica, Ficaria Vernon, mercurial perennis, Stellar holostea, Carex sylvatica, Phyteuma spicatum. The methods of work: The 1991, 2000, 2006, 2007, 2008 year meteorological and phenological indicators was selected by the autumn of ten 1x1 m size barrels, which is structured as follows, that the widest possible range of plant species employed in fixed space. Each year from spring to fall once a week has been register the growing species of heights, projectors coverage and phenological phases. Precipitation and temperature data has been analysed. Graphic depiction of data used in Microsoft Office XP software. The results of works: The results of this study show that each plant reacts differently to temperature and precipitation variation in accordance with their needs. Early vegetation and flowering date for each plant was 2007 and 2008. Strong correlations between the beginning... [to full text]

This thesis considers the entire automated speech recognition process and presents a standardised approach to LVCSR experimentation with HMMs. It also discusses various approaches to speaker adaptation such as MLLR and multiscale, and presents experimental results for cross­-task speaker adaptation. An analysis of training parameters and data sufficiency for reasonable system performance estimates are also included. It is found that Maximum Likelihood Linear Regression (MLLR) supervised adaptation can result in 6% reduction (absolute) in word error rate given only one minute of adaptation data, as compared with an unadapted model set trained on a different task. The unadapted system performed at 24% WER and the adapted system at 18% WER. This is achieved with only 4 to 7 adaptation classes per speaker, as generated from a regression tree.

This thesis considers the entire automated speech recognition process and presents a standardised approach to LVCSR experimentation with HMMs. It also discusses various approaches to speaker adaptation such as MLLR and multiscale, and presents experimental results for cross­-task speaker adaptation. An analysis of training parameters and data sufficiency for reasonable system performance estimates are also included. It is found that Maximum Likelihood Linear Regression (MLLR) supervised adaptation can result in 6% reduction (absolute) in word error rate given only one minute of adaptation data, as compared with an unadapted model set trained on a different task. The unadapted system performed at 24% WER and the adapted system at 18% WER. This is achieved with only 4 to 7 adaptation classes per speaker, as generated from a regression tree.

Two speech recognition methods: Dynamic Time Warping and Hidden Markov model based methods were investigated in this work To estimate efficiency of the methods, speaker dependent and speaker independent isolated word recognition experiments were performed. During experimental research it was determined that Dynamic Time Warping method is suitable only for speaker dependent speech recognition. Hidden Markov model based method is suitable for both – speaker dependent and speaker independent speech recognition.

One of the most common methods for isolated words recognition is based on Hidden Markov models. Speech signal can be considered as a sequence of successive parts of the signal with specific statistical parameters. Hidden Markov model corresponds to the statistical model with the final number of states, which may be useful for signals such as speech. HTK module is a software tools, which is mostly used to work with hidden Markov models.

This work describes development of a diff tool working with audio files containing general sound such as music, speech and other sounds. There are presented facts from different domains of science related to sound, such as psychoacoustics, speech recognition and automatic music genre categorisation. This paper also contains description of some diff algorithms and external tools needed for development of the goal application. Moreover, there is introduced design and implementation of the application, settings used for sound features extraction and evaluation of attained results.

In this dissertation, methods to synthesize ZnO are reported. First, zinc oxide nanoparticles were synthesized with small amounts of transition metal ions to create materials called dilute magnetic semiconductors (DMS). We employed a low temperature sol-gel method that produces ZnO nanoparticles of reproducible size and incorporates cobalt, nickel, and manganese ions into the nanoparticles. Conditions were controlled such that a range of amounts of Co, Ni, and Mn were incorporated. The incorporation was tracked by color changes in the white ZnO powder to blue for Co, green for Ni and yellow for Mn. XRD measurements showed the nanoparticles were on the order of 10 nm in diameter and had a wurtzite structure. Magnetic measurements showed a change from diamagnetic to paramagnetic behavior with increasing concentration of metal dopants. Second, formation of ZnO single crystal hexagonal prisms from a sol-gel method is presented. The method required water, zinc acetate, and ethanolamine to create a gel of zinc hydroxide and zinc hydroxide acetate, which upon heating formed single crystal hexagonal prisms. Characterization of the gel was done by XRD as well as XRD high temperature chamber (HTK) to determine the role of temperature in prism formation. SEM images showed hexagonal prisms were of uniform size (0.5 × 2 µm.) TEM and electron diffraction images showed a change from randomly oriented particles to an ordered single crystal after heating. Water and the acetate salt of zinc proved to be critical to prism formation. Lastly, we report absorption and fluorescence properties of synthesized oligothiophenes and oligothiophene-ruthenium complexes that are bound to CdSe nanoparticles. Their ability to act as sensitizers and charge transfer junctions was tested. It was found that fluorescence of CdSe nanoparticles was quenched when they were bound to the oligothiophenes, and that the fluorescence of the oligothiophenes was also quenched. The fluorescence lifetimes of the quenched species were shortened.