Voice recognition technologies appeared in the period of general device miniaturization, when all technologies were commonly integrated into one lust. There is no space for buttons and displays anymore. To have a good system of Lithuanian language recognition, a number of throughout researches must be implemented. Only after selecting the most efficient speech recognition scheme, we can proceed to the development of software adapted to the contemporary time. The aim of this paper is to determine, how efficient speech recognition is possible using neuron networks. MFCC and LPC coefficients were chosen as the parameters characterizing the phonemes. The paper attempts at the determination of the coefficients, which lead to the most efficient recognition of phonemes. For testing, programs PRAAT and MatLab were used. After implementing a number of phoneme recognition experiments in the research work, the results were obtained, which lead to the following conclusions: 1. In case of using neuron network for the recognition of isolated sounds and characterizing the phonemes by MFCC or LPC coefficients, the possibility of recognition does not exceed 90 per cent. It is not enough for quality recognition of Lithuanian speech. 2. In case of using MFCC coefficients, separate phonemes are recognized better than using LPC coefficients. The difference is about 15 per cent. 3. The advantage of LPC coefficients in comparison with MFCC is the curve of recognition possibility, which is more even... [to full text]

ČIULADIS, Paulius (2011) Diktoriaus atpažinimo tyrimas naudojant stacionarią fonemos dalį. Magistro baigiamojo darbo ataskaita. Kaunas: Vilniaus universitetas, Kauno humanitarinis fakultetas, Informatikos katedra. 48psl Magistro baigiamojo darbo tikslas - Nustatyti ar fonemos stacionarioji dalis turi būdingų konkrečiam diktoriui savybių, kurios leidžia jį identifikuoti arba nustatyti, kuriai grupei (vyrų ar moterų) priklauso. Siekiant tikslo darbe sprendžiami uzdaviniai: 1)Išanalizuoti „Diktoriaus atpažinimo naudojant stacionarią fonemos dalį, išskirtą kepsrinių koeficientų pagalba“ mokslinę temą; 2) Sudaryti ir paruošti diktorių atpažinimui reikalingų įrašų grupę; 3) Sukurti skirtingų diktorių fonemos staconarios dalies savybių išskyrimo ir palyginimo algoritmą; 4) Atlikti eksperimentą su paruoštais įrašais naudojant sukurtą algoritmą; 5) Išanalizuoti eksperimentų metu, pasiektus rezultatus. Išskiriant fonemos stacionarią dalį naudotas segmentavimo metodas. Iš signalo fonemų segmentų atrinktų tolimesnei analizei buvo išskiriami Melų dažnių kepstriniai koeficientai (MDKK). Atlikus numatytus uždavinius ir realizavus eksperimentą, diktoriaus atpažinimas pagal stacionarius fonemos fragmentus išskiriant kepstrinius koeficientus pirmo bandymo metu laido teisingi atpažinti diktorius vos 26,3 % visų atvejų, o pagal lytį net 91,8 % tikslumu. Pagal identiškame kontekste ištartų fonemų stacionarius fragmentus išskiriant kepstrinius koeficientus, antro bandymo metu laido teisingai... [toliau žr. visą tekstą] / ČIULADIS, Paulius (2011) Speaker Recognition Using Stationary Part of Phoneme. Master's thesis report. Kaunas: Vilnius University, Kaunas Faculty of Humanities, Department of Computer Science. 48psl Master's thesis is - Set a fixed part or phonemes specific to a particular speaker characteristics, which enables it to identify or determine which group (men or women) are. Towards the goal of this work, the following objectives: 1) analyze the broadcaster 's use of fixed recognition of phonemes, the isolated kepsrinių factors support "scientific topic, 2) Develop and prepare the necessary records announcers identification group, and 3) create different phonemes staconarios announcers of the nature of isolation and comparison of algorithms; 4) To perform the experiment with records prepared by using an algorithm, 5) to analyze the experiments, the results achieved. Distinguishing between phonemes used in the fixed part of the segmentation method. Phonemic segments of the signal selected for further analysis of the spectrum lies kepstriniai secreted factors (MDKK). After realizing the objectives set and the experiment narrator recognition by fragments of the stationary phonemes kepstrinius distinguishing factors of the first test to identify the correct speaker wire just 26.3% of all cases by sex and even the 91.8% accuracy. Under identical to pronounce the phonemes in the context of distinguishing between fixed fragments kepstrinius coefficients of the second test correctly... [to full text]