|
|
Speech Recognition For Selected Languages
Schmitt, Jan ; Karafiát, Martin (referee) ; Janda, Miloš (advisor)
This bachelor's thesis deals with recognition of continues speech for three languages - Bulgarian, Croatian and Swedish. There are described basics of speech processing and recognition methods like acoustic modeling using hidden Markov models and gaussian mixture models. Another aim of this work is preparing data for those languages from GlobalPhone database, so they may be used with speech recognition toolkits Kaldi and HTK. With data prepared there are several models trained and tested using Kaldi toolkit.
|
|
|
Voice recognition of standard PILOT-CONTROLLER control commands
Kufa, Tomáš ; Polách, Petr (referee) ; Honzík, Petr (advisor)
The subject of this graduation thesis is an application of speech recognition into ATC commands. The selection of methods and approaches to automatic recognition of ATC commands rises from detailed air traffic studies. By the reason that there is not any definite solution in such extensive field like speech recognition, this diploma work is focused just on speech recognizer based on comparison with templates (DTW). This recognizor is in this thesis realized and compared with freely accessible HTK system from Cambrige University based on statistic methods making use of Hidden Markov models. The usage propriety of both methods is verified by practical testing and results evaluation.
|
|
|
Character recognition in the soundtrack with SOM
Malásek, Jan ; Honzík, Petr (referee) ; Honzík, Petr (referee) ; Pohl, Jan (advisor)
This bachelor´s thesis describes a history of neural networks evolution and their using in speech recognition systems and shows problems with working and learning neural networks. It presents three chosen systems for speech recognition including their evaluation in experiments, their advantages and disadvantages. It is also about human speech characteristics and systems of its recognition. The last part is focused on frequency spectrums of different types of vowels and gives instructions for programming neural networks using MATLAB.
|
|
|
Speech Recognition (digit)
Kantar, Martin ; Minář, Petr (referee) ; Matoušek, Radomil (advisor)
The aim of this diploma thesis is to explain what speech is and what are its constituents. I mention commonly used methods which are used for preparation of signals which we use for recognition. Schematic examples show principles of current recognizers of speech, their advantages and disadvantages. I made speech recognition program for 0-9 numerals in Matlab for neural nets learning.
|
|
|
Recognition of Multi-Talker Overlapping Speech Using Neural Networks
Hradil, Jaromír ; Švec, Ján (referee) ; Žmolíková, Kateřina (advisor)
Tato práce se zabývá rozpoznáváním řeči překrývajících se řečníků pomocí neuronové sítě. Zkoumá problém rozpoznávání řečí od vícero řečníků a způsoby, jimiž se tento daný problém řeší. Jedná se konkrétně o aplikaci kromě tradičních komponentů jako konvoluční neuronové sítě, LSTM atd. také speciálních komponentů: attention mechanismus a gated konvoluce. A dále také aplikace techniky zvanou permutation invariant training. Součástí této práce je aplikování těchto přístupů na přidělená trénovací data, která jsou tvořena uměle vytvořenými směsmi dvou řečníků předčítající články z Wall Street Journal. Dalším krokem bylo natrénování příslušných architektur používající kombinující prvky zmíněné nahoře. Modely v této práci nahrazují akustický model. Jednalo se o dvě architektury užívající různé typy attention mechanismu a o jednu bez něj. Experimenty ukázaly, že architektury užívající attention mechanismus v tomto typu úlohy něpřekonaly tradičnější architekturu s užitím gated konvolucí. Přesto ale ukázaly potenciál.
|
|
|
Voice Controlled Calculator
Pavelek, Ota ; Szőke, Igor (referee) ; Grézl, František (advisor)
This thesis describes implementation of calculator, which can be controlled by both voice and normal way. Speech recognizing is realized with BSCORE library and has recognition network restricted to words needed by calculator. When the recognition is done, recognized sentence is transformed to expression and shown to user, so it can be corrected (especially in case of erroneous recognition). Expression is evaluated on user's request. The purpose of voice control is to make usage of calculator more effective and accesible for handicapped users.
|
|
|
Detection of speech disorders
Struhař, Michal ; Rajmic, Pavel (referee) ; Sysel, Petr (advisor)
This thesis deals with detection of speech disorders. One of the aims of this thesis is choosing suitable parameterization: short-time energy, zero-crossing rate, linear predictive analysis, perceptual linear predictive analysis, RASTA method, cepstral analysis and mel-frequency cepstral coefficient can be chosed for detections. Next aim is construction of detector of speech disorders based on DTW (Dynamic Time Warping) and artificial neuron network. Single detection proceeds on the base of collected tokens from chosen analysis and phonetic transcription of speech. Analyses, detector and phonetic transcription of Czech language are implemented in simulation environment of MATLAB.
|
|
|
Vizualization of Outputs from Speech Technologies for Contact Centers
Zhezhela, Oleksandr ; Szőke, Igor (referee) ; Schwarz, Petr (advisor)
The thesis is aimed on visualisation of data mined by speech processing technologies. Some methods speech data extraction were studied and technologies for this task were analysed. The variety of meta data that can be mined from speech was defined. Were also examined existing standards and processes of call centres. Some requirements for the user interface were gathered and analysed. On that basis and after communication with call centre employees there was defined and implemented a concept for speech data visualization. Gained solutions were integrated into Speech Analytics Server (SPAS).
|
|
|
Speech Recognition Algorithms in FPGA/DSP
Urbiš, Oldřich ; Herout, Adam (referee) ; Szőke, Igor (advisor)
This master's thesis deals with design of speech recognition algorithms with consideration of target technology, which is platform combinating digital signal processing and field programmable gate array. Algorithms for speech recognition includes: feature extraction of Melfrequency cepstral coefficients, hidden Markov models and their evaluation by Viterbi algorithm.
|
|
|
Language Modeling for Spech Recognition in Czech
Mikolov, Tomáš ; Černocký, Jan (referee) ; Smrž, Pavel (advisor)
This work concerns the problematic of language modeling in automatic speech recognition. Currently widely used techniques for advanced language modeling based on statistical approach are described in the first part of work - class based language models, factored language models and neural network based language models. In the next section, implementation of neural network based language model is described. Results obtained on "Pražský mluvený korpus" and "Brněnský mluvený korpus" corpora (1 170 000 words) are reported, with perplexity reduction around 20%. Also, results obtained after rescoring N-best lists with spontaneous speech are reported, with absolute improvement in accuracy by more than 1%. In the conclusion, possible uses of the work are mentioned, along with possible extensions in the future. Finally, main weaknesses of current statistical language modeling techniques are described.
|
guest ::
