|
|
Robust Speech Activity Detection
Popková, Anna ; Plchot, Oldřich (referee) ; Matějka, Pavel (advisor)
The aim of this work is to design and create a robust speech activity detector that is able to detect speech in different languages, in a noise environment and with music on background. I decided to solve this problem by using a neural network as a classification model that assigns one of the four possible classes - silence, speech, music, or noise to the input of audio recording. The resulting tool is able to detect the speech in at least 12 languages. Speech with musical background up to 88 % accuracy and system success on noisy data reaches from 84 % (5 dB SNR) to 88 % (20 dB SNR). This tool can be used for speech activity detection in various research areas of speech processing. The main contribution is the elimination of music, which when not eliminated, significantly increases the error rate of systems for speaker identification or speech recognition.
|
|
|
Agreements and Disagreements between Automatic and Human Speaker Recognition
Valenta, Jakub ; Matějka, Pavel (referee) ; Rohdin, Johan Andréas (advisor)
Tato práce se zabývá problémem rozpoznáváním mluvčího. Uvedený pojem je definován a doplněn o jednotlivé metody, které s ním souvisí. Cílem práce je poukázat na shody a rozdíly mezi lidským a automatickým procesem rozpoznávání mluvčího. V úvodu práce jsou popsány teoretické poznatky z obou zmíněných oblastí, tj. na jaké aspekty lidské řeči se zaměřuje člověk, resp. automatický systém. Následně je provedeno několik experimentů, které mají za úkol srovnat tyto dvě metody. Tyto experimenty jsou vyhodnoceny tak, že je možné pozorovat, které testovací úlohy dokáže lépe vyřešit člověk, aby následně bylo možné tyto poznatky použít ke zlepšení funkce automatického systému. V závěru práce je takovýto návrh na zlepšení automatického systému předveden a otestován. Testování proběhlo úspěšně a byla zaznamenána vyšší přesnost při vyhodnocování. Takový výsledek tedy může být užitý v dalších výzkumech a umožnit tak další vývoj v oblasti automatického rozpoznávání mluvčích.
|
|
|
Acoustic Scene Classification from Speech
Grepl, Filip ; Beneš, Karel (referee) ; Matějka, Pavel (advisor)
This thesis deals with creating a system whose task is to recognize what type of location the recording was created at by analyzing the audio signal. The classifier is based on a multi-layer, fully connected neural network. The topology of the neural network is based on the baseline system provided for the DCASE competition. A dataset from this competition is also used for training and evaluating the neural network. The experiments are performed in particular with the representation of the properties of the audio records and with the format of the input data of the neural network. For this purpose, Mel-filter bank, block Mel-filter bank and MFCC flags are used. The experiments performed in this thesis brought a classification accuracy increased by 6.5 % compared to the baseline system of DCASE. Overall system success rate reached 67.5 %.
|
|
|
Intelligent Meeting Room Controlled by Voice
Bauer, Jan ; Matějka, Pavel (referee) ; Schwarz, Petr (advisor)
The aim of the thesis is to design and create system for intelligent room controlled by voice. The solution is based on Phonexia Speech Engine developed by Phonexia. The system runs on Raspberry Pi. The core functionality of the system is implemented in Python. The resulting solution is certainly interesting and with some updates may become intelligent assistant for meettings.
|
|
|
Acoustic Scene Classification from Speech
Dobrotka, Matúš ; Glembek, Ondřej (referee) ; Matějka, Pavel (advisor)
The topic of this thesis is an audio recording classification with 15 different acoustic scene classes that represent common scenes and places where people are situated on a regular basis. The thesis describes 2 approaches based on GMM and i-vectors and a fusion of the both approaches. The score of the best GMM system which was evaluated on the evaluation dataset of the DCASE Challenge is 60.4%. The best i-vector system's score is 68.4%. The fusion of the GMM system and the best i-vector system achieves score of 69.3%, which would lead to the 20th place in the all systems ranking of the DCASE 2017 Challenge (among 98 submitted systems from all over the world).
|
|
|
Designing and testing of new metal nanosubstrates for biomolecular sensors based on surface-enhanced Raman scattering (SERS) spectroscopy
Peksa, Vlastimil ; Procházka, Marek (advisor) ; Matějka, Pavel (referee) ; Richter, Ivan (referee)
Title: Designing and testing of new metal nanosubstrates for biomolecular sensors based on surface-enhanced Raman scattering (SERS) spectroscopy Author: Vlastimil Peksa Department: Institute of Physics of Charles University Supervisor: doc. RNDr. Marek Procházka, Ph.D., Institute of Physics of Charles University Abstract: This experimental methodical work was aimed at the optimization of selected gold and silver substrates and their use in construction of SERS-based biosensors, including following practical application. Several types of substrates, fabricated via a combination of bottom-up techniques on solid surfaces, were tested. The properties of these substrates were examined with probe molecules, namely methylene blue, porphyrins and tryptophan, on a confocal Raman microspectrometer. Obtained findings about the influence of analyte application, objective focusing and internal intensity standard were exploited for optimization of measurement procedures with regard to sensitivity, accuracy and reproducibility. A method for quantitative detection of food dye azorubine (E 122) in commercially available drinks was developed, based on these findings. Its results have shown its potential as a pre-scan method for field application and preliminary testing. Keywords: Metal nanosubstrates, biomolecules,...
|
|
|
Codec Detection from Speech
Jon, Josef ; Matějka, Pavel (referee) ; Černocký, Jan (advisor)
Tato práce se zabývá detekcí kodeků z komprimovaného řečového signálu. Cílem bylo zjistit, jaké charakteristiky rozlišují jednotlivé kodeky a následně vytvořit prostředí vhodné pro experimenty s různými typy a konfiguracemi klasifikátorů. Použity byly Support vector machines a především neuronové sítě, které byly vytvořeny pomocí nástroje Keras. Hlavním přínosem této práce je experimentální část, ve které je analyzován vliv různých parametrů neuronové sítě. Po nalezení nejvhodnější kombinace parametrů dosáhla síť přesnosti klasifikace přes 98% na testovací sadě obsahující data z 6 kodeků.
|
|
|
Magnetic circular dichroism and aromatic compounds
Štěpánek, Petr ; Bouř, Petr (advisor) ; Matějka, Pavel (referee) ; Srnec, Martin (referee)
Title: Magnetic circular dichroism and aromatic compounds Author: Petr Štěpánek Department/Institute: Institute of Organic Chemistry and Biochemistry AS CR, v.v.i. Supervisor: prof. RNDr. Petr Bouř, DSc., Institute of Organic Chemistry and Biochemistry AS CR, v.v.i. Abstract: The thesis presents a series of studies concerning magnetic circular dichroism (MCD), a spectroscopic method, which experienced an intense theo- retical development in the recent years. New computational codes opened possi- bilities to calculate MCD spectra of larger and more varied molecules than was possible in the past. In the presented studies, we took the advantage of the new computational codes to broaden the possible span of applications of the MCD technique. As an example, we present MCD as a method useful for obtaining information about the structure of fullerenes. We also studied the influence of the molecular conformation and the explicit and implicit solvent models on the MCD spectra of aromatic amino acids using the newly implemented alterna- tive computational protocol based on sum-over-states calculations. We have also theoretically predicted spectra of the nuclear spin circular dichroism (NSCD), a potential new high-resolution spectroscopy. Keywords: magnetic circular dichroism, quantum-chemical calculations, density...
|
|
|
Raman optical activity of biomolecules: From simple models to complex systems
Pazderková, Markéta ; Baumruk, Vladimír (advisor) ; Kapitán, Josef (referee) ; Matějka, Pavel (referee)
The aim of the thesis is to utilize Raman optical activity (ROA) to get unique information on peptide/protein conformation, which is otherwise difficult or even impossible to obtain. We have focused on investigation of amide and disulfide groups. Utilizing tailor-made model structures (rigid tricyclic spirodilactams with two interacting nonplanar amide groups), special model peptides and even biologically active molecules (neurohypophyseal hormones and their agonistic and antagonistic analogs, antimicrobial peptide lasiocepsin and its analogs having different disulfide pattern) we have traced specific spectral manifestation of nonplanar amides and disulfides. ROA results were supplemented by data obtained by complementary chiroptical methods - electronic (including vacuum UV - SRCD) and vibrational circular dichroism. When used in a concerted fashion, these techniques provide complex information on peptide/protein secondary structure. Where possible, experimental chiroptical data were compared to ab initio calculations. In chiroptical spectra we have found and interpreted signals reflecting nonplanarity of the amide group. Moreover, in ROA spectra we have identified signals due to S-S stretching vibrations which seem to reflect sense of the disulfide group torsion.
|
|
|
Synthesis and study of ceramide analogues
Školová, Barbora ; Vávrová, Kateřina (advisor) ; Matějka, Pavel (referee) ; Šklubalová, Zdeňka (referee)
Charles University in Prague, Faculty of Pharmacy in Hradec Králové Department of inorganic and organic chemistry Candidate: Mgr. Barbora Školová Supervisor: doc. PharmDr. Kateřina Vávrová, Ph.D. Title of doctoral thesis: Synthesis and study of ceramide analogues Ceramides (Cer) together with free fatty acids and cholesterol form intercellular lamellar space of the uppermost skin layer (stratum corneum, SC). This lipid matrix presents the proper skin barrier - it protects organism against outer environment and keeps its homeostasis as well. The composition and arrangement of the intercellular lipid matrix is essential for the skin barrier properties. In spite of this fact, the organization of SC lipids at the molecular level is not fully elucidated. The aim of this work is to contribute to more detailed insight into organization of skin barrier using SC lipid model membranes, in particular to find the structural parameters in ceramide molecules, which play a role in the maintenance of the skin barrier. The SC model membranes present a useful tool to study SC lipids, they mimic physiological or pathological conditions in the skin. The obtained results should be used in diagnosing or therapy of skin diseases characteristic with impaired barrier function, such as atopic dermatitis or psoriasis....
|
guest ::
RSS feed.