Search

The new 3D motion capture data corpus expands the portfolio of existing language resources by a corpus of 18 hours of Czech sign language. This helps alleviate the current problem, which is a critical lack of quality data necessary for research and subsequent deployment of machine learning techniques in this area. We currently provide the largest collection of annotated sign language recordings acquired by state-of-the-art 3D human body recording technology for the successful future deployment of communication technologies, especially machine translation and sign language synthesis.

In this paper, it is discussed the problem of long-term visual localization with a using of the Aachen Day-Night dataset. Our experiments confirmed that carefully fine-tuning parameters of the Hierarchical Localization method can lead to enhance the visual localization accuracy. Next, our experiments show that it is possible to find an image’s area that does not add any valuable information in long-term visual localization and can be removed without losing the localization accuracy. The approach of using the method of semantic segmentation for preprocessing helped to achieve comparable state-of-the-art results in the Aachen Day-Night dataset.

In this paper, we focus on the issue of speaker diarization of data from a real call center. We have previously proposed a specialized solution to the problem, which employed additional knowledge about the identities of the phone operators (in our case, the language counselors from the Language Consulting Center), thus improving performance over the baseline. But a recent end-to-end diarization method, EEND, has since proven very successful on other data and was shown to surpass the previous state of the art in the field. Thus, we chose to compare this new method with our own previous approach. Using an existing implementation of the EEND method (adapted using a small amount of the target data from the Language Consulting Center), we successfully surpass the performance of our previous approach (17.42% vs. 19.39% DER), without the need for any additional information about speaker identities. The majority of the remaining diarization error of the EEND system is due to incorrect decisions between speech and silence, rather than speaker confusion. For comparison, we also show the results of a more standard diarization approach, represented by the method used in the Kaldi toolkit.

Tento článek se zabývá úkolem extrakce informací ze strukturovaného dokumentu skenovaného běžným kancelářským skenerem. Zkoumá přístupy zpracování naskenovaných papírových dokumentů a extrakci hledaných informací, jako jsou jména, adresy, data a další číselné hodnoty. Představujeme návrh systému rozděleného do čtyř po sobě jdoucích modulů: předzpracování, optické rozpoznávání znaků, extrakce informací pomocí databáze a extrakce informací bez databáze. V modulu předzpracování jsou představeny dvě základní techniky - zlepšení kvality obrazu a odklonění obrazu. Řešení optického rozpoznávání znaků a přístupy k extrakci informací jsou porovnávány pomocí výkonu celého systému. Nejlepší výkon extrakce informací s databází byl získán algoritmem Locality-sensitive Hashing. This paper deals with the task of information extraction from a structured document scanned by an ordinary office scanner device. It explores the processing pipeline from scanned paper documents to the extraction of searched information such as names, addresses, dates, and other numerical values. We propose system design decomposed into four consecutive modules: preprocessing, optical character recognition, information extraction with a database, and information extraction without a database. In the preprocessing module, two essential techniques are presented – image quality improvement and image deskewing. Optical Character Recognition solutions and approaches to information extraction are compared using the whole system performance. The best performance of information extraction with the database was obtained by the Locality-sensitive Hashing algorithm.

V tomto článku popisujeme diarizaci mono telefonních dat z Jazykové poradny Ústavu pro jazyk český. Náš navrhovaný přístup k diarizaci využívá informace o identitě jednoho z účastníků hovoru. V klasickém přístupu k diarizaci nahrazujeme shlukování x-vektorů identifikací řečníka. In this paper, we describe a diarization of mono channel telephone recordings from The Language Consulting Center providing the Czech language consultancy service. In our proposed approach to a diarization, we use information about the known identity of one speaker (the language counsellor) acquired from the text transcription at the beginning of the conversation. In the state-of-the-art diarization based on the x-vectors clustering, we replace the clustering step by the identification of each segment of the recording against the counsellor’s identity x-vector and the general x-vector model that represents the client. Our proposed diarization without resegmentation step can be used as an online approach. Because of the uniqueness of our data, we compare our results with the Kaldi diarization as the baseline system.

V tomto článku popisujeme diarizaci archivu Jazykové poradny vznikajícím v rámci projektu "Zpřístupnění dotazů jazykové poradny v lingvisticky strukturované databázi". Jedna část tohoto archivu je nahraná pouze v mono kvalit, naším úkolem je proto rozdělit data pomocí diarizace. Náš přístup využívá informace o identitě jazykového poradce získané z přepisu jeho představení na začátku každého z hovorů. Protože naše data jsou jedinenčná, pro porovnání uvádíme také výsledky dostupného systému diarizace Kaldi. In this paper, we describe a diarization of the archive data from the project “Access to a Linguistically Structured Database of Enquiries from the Language Consulting Center”. This project is attempting to provide improved access to the large archives of the Czech language of mainly telephone conversations collected continuously by The Language Consulting Center. One part of this archives contains mono recordings, where the data of the client and the language counsellor are mixed in one channel. In our proposed approach to a diarization, we used the information about the identity of the language counsellor acquired from the text transcription on the beginning of the conversation. For the initial stage of the diarization, our system based on clustering the x-vectors was adopted. The resegmentation step is used for refining the boundaries of speaker changes by the pre-trained Gaussian mixture model of the counsellor. Because of the uniqueness of our data, we compared our results with the Kaldi diarization as the baseline system.

In this paper, we describe the initial stages of the project “Access to a Linguistically Structured Database of Enquiries from the Language Consulting Center”. This project is attempting to provide an improved access to the large archives of mainly telephone conversations collected continuously by the Institute of the Czech Language. The main goal is to open up the unique Czech data acquired from the queries to the Language Consulting Center and to build the semi-automatic system that will facilitate searching and categorizing of these queries. For this purpose, the Automatic Speech Recognizer (ASR) and the language processing methods are being designed. The vocabulary used in such queries contains many unusual words unlike the common speech (e.g. linguistic terms). In order to train the ASR system, it is necessary to manually transcribe a large amount of speech data, identify the appropriate vocabulary, and obtain relevant text for language modeling purposes. In this paper, the proposed telephone system for recording the new data and the baseline speech recognition on these data is described. The first experiments with the topic detection on these data aimed at discovering what can be found in them and also how to preprocess them is also described.

In this paper, we propose a speaker change detection system based on lexical information from the transcribed speech. For this purpose, we applied a recurrent neural network to decide if there is an end of an utterance at the end of a spoken word. Our motivation is to use the transcription of the conversation as an additional feature for a speaker diarization system to refine the segmentation step to achieve better accuracy of the whole diarization system. We compare the proposed speaker change detection system based on transcription (text) with our previous system based on information from spectrogram (audio) and combine these two modalities to improve the results of diarization. We cut the conversation into segments according to the detected changes and represent them by an i-vector. We conducted experiments on the English part of the CallHome corpus. The results indicate improvement in speaker change detection (by 0.5% relatively) and also in speaker diarization (by 1% relatively) when both modalities are used.

The marine n-3 fatty acids eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) exert numerous beneficial effects on health, but their potency to defend against development of peripheral insulin resistance of healthy person with overweight remains poorly characterized. We aimed to evaluate the effect of a combination intervention using EPA + DHA and the lifestyle modification (LSM) in women with overweight.In a parallel-group, three-arm, randomized trial (UMIN Clinical Trials Registry - R000031131), 34 women were assigned to a 12-week-intervention using corn oil (1.5 g/day; placebo); LSM and corn oil (1.5 g/day; LSM); or LSM and EPA + DHA concentrate (1.5 g/day, containing ~ 0.6 g EPA + DHA; LSMn-3). At baseline and after intervention, anthropometric measurements including bioelectrical impedance analysis, spiroergometry, 24-hours dietary recall, and various metabolic markers, adiponectin and cytokines were evaluated in serum using standard procedures. Data from 29 women were used for the final evaluation. Wilcoxon two-sided rank-sum test was used to inspect the differences between LSM and LSMn-3, and placebo groups, with a p-value of ≤ 0.05. All computations were performed with MATLAB Statistics Toolbox.In comparison with placebo, LSM and LSMn-3 decreased body weight, waist circumference, and body fat, and increased VOBesides preventing a modest negative effect of LSM on IL6 and adiponectin level, the combination of LSM and EPA + DHA supplementation could be probably used to improve the functional capacity of adipose tissue in women with overweight.

Chromium is required for a normal insulin function, and low levels have been linked with insulin resistance. The aim of this study was to follow the effect of chromium supplementation on fasting plasma glucose (FPG), glycated haemoglobin (HbA1c) and serum lipids in patients with type 2 diabetes mellitus (DM2) on insulin therapy. Eleven randomly selected patients with DM2 on insulin therapy were supplemented with a daily dose of 100 μg chromium yeast for the first supplementation period of 2 weeks. In the second supplementation period, the chromium dose was doubled and continued for the next 6 weeks. The third phase was a 6-week washout period. After each period, the levels of FPG and HbA1c were compared with the corresponding values at the end of the previous period. Serum triglycerides, total HDL and LDL cholesterol values after supplementation were compared with the baseline values. FPG decreased significantly after the first period of chromium supplementation (p < 0.001), and a tendency to a further reduction was observed after the second supplementation period. Similarly, HbA1c decreased significantly in both periods (p < 0.02 and p < 0.002, respectively). Eight weeks after withdrawal of chromium supplementation, both FPG and HbA1c levels returned to their pre-intervention values. The serum lipid concentrations were not significantly influenced by chromium supplementation. Chromium supplementation could be beneficial in patients with DM2 treated with insulin, most likely due to lowered insulin resistance leading to improved glucose tolerance. This finding needs to be confirmed in a larger study.

In this paper, we compare four methods of feature extraction for finding similar images. We created our own dataset which consists of 34500 colour images of various types of footwear and other accessories. We describe the methods that we experimented with and present results achieved with all of them. We presented the results to 23 people who then rated the performance of these methods. The best–rated method is selected for further research.

It is assumed that human exposure to phthalates may be associated with adverse health effects. The indicators of urinary phthalate metabolite concentrations in healthy adults are limited. In this study, the phthalate metabolites concentrations were detected from 24-h urine collection in non-obese Czech adults (n = 201). Each participant filled in an 80-item questionnaire (FANTOM-SQ 2013) regarding the outdoor and indoor sources of phthalates, the use of personal care products and food intake sources. The concentrations of 15 phthalates metabolites were analysed following enzymatic cleavage of the glucuronide using ultra-high-performance liquid chromatography-electrospray ionisation tandem mass spectrometry (UHPLC-ESI-MS/MS). The indicators of chronic or acute exposure phthalate-containing materials were identified. It is shown that higher fruit consumption was positively and significantly associated with a higher level of total 15 urinary phthalates biomarkers (p

The aim of this study was to reveal whether accumulation of the persistent organic pollutants (POPs), especially polychlorinated biphenyl (2,2',4,4',5,5'-hexachlorobiphenyl, PCB 153), affects plasma levels of adiponectin in obese patients. The study was designed as a longitudinal intervention trial with a control group, where 27 obese women (body mass index (BMI)30 kg/m(2); age 21-74 years) were studied before (OB) and after (OB-LCD) a 3-month low-calorie-diet intervention (LCD; 5 MJ daily). As the control group, 9 female volunteers without LCD intervention were used (C; BMI=19-25 kg/m(2); age 21-64 years). Plasma levels of PCB 153 were measured by high-resolution gas chromatography with electron capture detection; total adiponectin and insulin plasma levels were quantified by immunoassays; and adiponectin multimeric complexes were quantified by immunoblotting. Plasma levels of total adiponectin, high and medium molecular weight multimers significantly negatively correlated with plasma levels of PCB 153 in OB, but not in C or in OB-LCD, whereas the LCD intervention lowered BMI by 3.3+/-3.0 kg/m(2). Our results may suggest suppression of adiponectin by PCB 153 in obese women under non-energy-restrictive regime, which may contribute to the known association of PCB 153 and other POPs with type 2 diabetes.

In this paper, we compare the performance of image keypoints detectors and descriptors on well known Oxford dataset. We use evaluation criteria which were presented by Mikolajczyk et al. [12, 13]. We created most of the possible combinations of keypoint detector and descriptor, but in this paper, we present only selected pairs. The best performing detector and descriptor pair are selected for future research, mainly with the focus on augmented reality.

INTRODUCTIONPersistent organochlorine pollutants (POPs), organochlorine pesticides and polychlorinated biphenyls (PCBs) are synthetic chemicals that have been released into the environment in hundreds of metric tons for many years and have been used in diverse applications such as industry and agriculture. Although they were banned in most countries, their residues are still widespread in the environment. Because of their high lipophilicity and structural stability, they still can be found accumulated in biological systems. They show a carcinogenic and an endocrine disruption potential. Recently, based on epidemiological studies, there has been concern about their speculated association with metabolic disturbances like diabetes mellitus type 2, rheumatoid arthritis, disturbances of cognitive function, etc. Contaminated food, especially fatty fish, meat and milk products, is still a main source of the exposure of general population (1).Among chlorinated pesticides, p,p'-DDT has been used most extensively. Its metabolite p,p'-DDE together with hexachlorocyclobenzene (HCB) are, at the present day, the most frequently prevalent residues in the adipose tissue of humans (2).PCBs form a subgroup of halogenated aromatic hydrocarbons. There are theoretically 209 single congeners. They have been used, to a great extent, in diverse technical applications from the 1930s to the 1970s (3). Despite production bans in many industrial countries in the 1970s and 1980s, about 113 different PCB congeners were still detectable in the environment in 1998 (4). Their concentration in human tissue is generally monitored by using examinations of human breast milk and data from Europe show their decreased concentration. 'Non-dioxin-like' PCB 153, 138, and 180 are the main congeners which show the highest concentration and are most frequently found in breast milk (5). In the former Czechoslovakia, PCB was marketed as technical mixtures under the trade name "Delor" produced in the amount of about 21,500 tons in Eastern Slovakia from 1959 to 1984, when the production was finally abolished (6). About 11,600 tons were used inside the former Czechoslovakia (7). Therefore, the industrial areas of the Czech Republic have one of the highest PCB burdens of all European countries (8).After human exposure, POPs are accumulated in the body, especially in the adipose tissue. They have relatively long halflives before disappearing from the tissue or from the body. The variability of the stored amount depends on dietary exposure (food habits, and the quality and quantity of POPs in available food) and on individual disposition to store these substances (genetic traits and size of adipose tissue and its dynamic changes) (9). Their distribution in the body is affected by changes in body weight. Weight loss resulted in an acute increase of their concentration in reduced adipose tissue, brain, and serum/plasma without changing the total amount, as it was documented in case of hexachlorocyclohexane (HCH) (10). However, dramatic weight loss, which occurred as a consequence of bariatric surgery treatment, showed not only an increased level of serum and adipose tissue POPs concentrations together with a reduced amount of white adipose tissue, but a decreased total body burden for POPs after 12 months post-surgery as well (11). The objective of our study was to analyze in five year follow up study the development of plasma POPs levels in obese women in relationship to their weight loss success, caused by restrictive diet regimen.MATERIALS AND METHODSTwenty obese adult women, aged 25-73 years (BMI > 30 kg/ m2) were randomly selected from the patients of the Centre for Obesity Treatment (Faculty Hospital in Pilsen, Czech Republic) during their first visit in 2006. Informed consent was obtained from all patients involved in the study. Anthropometry: body weight (BW) and height were measured, and a plasma analysis of 7 POPs (p,p'-DDE, p,p'-DDT, HCB, HCH s, PCB 153, 138, 180) was performed in all the subjects. …

Nutrition plays an important role in the health and development of children. Therefore, it is necessary to make the correct recommendations as well as knowing the real state of nutrition of the children involved. In the scope of prospective longitudinal monitoring of 200 children, randomly selected at birth from 2 regions of the Czech Republic, an examination of their dietary intake and food composition in the year 2000/2001 during their 14-25 months of age (mean 19.2±2.7 months) was performed. A standardised food frequency questionnaire was administered to the whole sample (response rate through mothers was 98%). Moreover, every second mother who has not yet breastfed the child was asked for a 24-h dietary weight record of her child (response rate 79%). Twelve percent of the children were long term breastfed. The “Food pyramid”of sample was created by the following daily portions: milk and dairy (3.5), other protein-rich food (1.0), fruits and vegetables (4.2) and cereals (2.5). The energy intake of non-...

This paper demonstrates how usual feature extraction methods such as the PLP can be successfully replaced by a neural network and how signal processing methods such as mean normalization, variance normalization and delta coefficients can be successfully utilized when a NN-based feature extraction and a NN-based acoustic model are used simultaneously. The importance of the deep NNs is also investigated. The system performance was evaluated on the British English speech corpus WSJCAM0.

In this paper, we present a new NN/HMM speech recognition system with a NN-base acoustic model and RNN-based language model. The employed neural-network-based acoustic model computes posteriors for states of context-dependent acoustic units. A recurrent neural network with the maximum entropy extension was used as a language model. This hybrid NN/HMM system was compared with our previous hybrid NN/HMM system equipped with a standard n-gram language model. In our experiments, we also compared it to a standard GMM/HMM system. The system performance was evaluated on the British English speech corpus and compared with some previous work.

The continuous skip-gram model is an efficient algorithm for learning quality distributed vector representations that are able to capture a large number of syntactic and semantic word relationships. Artificial neural networks have become the state-of-the-art in the task of language modelling whereas Long-Short Term Memory (LSTM) networks seem to be efficient training algorithm.

The aim of this work is to propose a refinement of the shift-MLLR (shift Maximum Likelihood Linear Regression) adaptation of an acoustics model in the case of limited amount of adaptation data, which can lead to ill-conditioned transformations matrices. We try to suppress the influence of badly estimated transformation parameters utilizing the Artificial Neural Network (ANN), especially Convolutional Neural Network (CNN) with bottleneck layer on the end. The badly estimated shift-MLLR transformation is propagated through an ANN (suitably trained beforehand), and the output of the net is used as the new refined transformation. To train the ANN the well and the badly conditioned shift-MLLR transformations are used as outputs and inputs of ANN, respectively.

Artificial neural networks have become state-of-the-art in the task of language modelling on a small corpora. While feed-forward networks are able to take into account only a fixed context length to predict the next word, recurrent neural networks (RNN) can take advantage of all previous words. Due the difficulties in training of RNN, the way could be in using Long Short Term Memory (LSTM) neural network architecture.

The worst problem the adaptation is dealing with is the lack of adaptation data. This work focuses on the feature Maximum Likelihood Linear Regression (fMLLR) adaptation where the number of free parameters to be estimated significantly decreases in comparison with other adaptation methods. However, the number of free parameters of fMLLR transform is still too high to be estimated properly when dealing with extremely small data sets. We described and compared various methods used to avoid this problem, namely the initialization of the fMLLR transform and a linear combination of basis matrices varying in the choice of the basis estimation (eigen decomposition, factor analysis, independent component analysis and maximum likelihood estimation). Initialization methods compensate the absence of the test speaker’s data utilizing other suitable data. Methods using linear combination of basis matrices reduce the number of estimated fMLLR parameters to a smaller number of weights to be estimated. Experiments are aimed to compare results of proposed basis and initialization methods.

In this paper we investigate whether a combination of statistical, neural network and cache language models can outperform a basic statistical model. These models have been developed, tested and exploited for a Czech spontaneous speech data, which is very different from common written Czech and is specified by a small set of the data available and high inflection of the words. As a baseline model we used a trigram model and after its training several cache models interpolated with the baseline model have been tested and measured on a perplexity. Finally, an evaluation of the model with the lowest perplexity has been performed on speech recordings of phone calls.

One of the most utilized adaptation techniques is the feature Maximum Likelihood Linear Regression (fMLLR). In comparison with other adaptation methods the number of free parameters to be estimated significantly decreases. Thus, the method is well suited for situations with small amount of adaptation data. However, fMLLR still fails in situations with extremely small data sets. Such situations can be solved through proper initialization of fMLLR estimation adding some a-priori information. In this paper a novel approach is proposed solving the problem of fMLLR initialization involving statistics from speakers acoustically close to the speaker to be adapted. Proposed initialization suitably substitutes missing adaptation data with similar data from a training database, fMLLR estimation becomes well-conditioned, and the accuracy of the recognition system increases even in situations with extremely small data sets.

The work presented in this paper deals with the text normalization for highly inflectional languages. This paper is focused on abbreviation expansion and likewise on numerals normalization. Our text normalization system does not use any explicit parser or part-of-speech tagger and thus it can be called lightly supervised. The standard rule-based text normalization method is compared with the proposed statistical-based one in the task of expansion of Czech abbreviations.

The main idea of a priori machine learning is to apply a machine learning method on a machine learning problem itself. We call it "a priori" because the processed data set does not originate from any measurement or other observation. Machine learning which deals with any observation is called "posterior". The paper describes how posterior machine learning can be modified by a priori machine learning. A priori and posterior machine learning algorithms are proposed for artificial neural network training and are tested in the task of audio-visual phoneme classification.

This paper describes ANN based posterior estimates and their application to speech recognition. We replaced the standard back-propagation with the L-BFGS quasi-Newton method. We have focused only on posterior based feature vector extraction. Our goal was a feature vector dimension reduction. Thus we designed three posterior transforms to space with dimensionality 1 or 2. The designed transforms were tested on the SpeechDat-East corpus. We also applied the introduced method on a Czech audio-visual corpus. In both cases the methods leads to significant word error rate decrease.

This paper deals with robust estimations of data statistics used for the adaptation. The statistics are accumulated before the adaptation process from available adaptation data. In general, only small amount of adaptation data is assumed. These data are often corrupted by noise, channel, they do not contain only clean speech. Also, when training Hidden Markov Models (HMM) several assumptions are made that could not have been fulfilled in the praxis, etc. Therefore, we described several techniques that aim to make the adaptation as robust as possible in order to increase the accuracy of the adapted system. One of the methods consists in initialization of the adaptation statistics in order to prevent ill-conditioned transformation matrices. Another problem arises when an acoustic feature is assigned to an improper HMM state even if the reference transcription is available. Such situations can occur because of the forced alignment process used to align frames to states. Thus, it is quite handy to accumulate data statistic utilizing only reliable frames (in the sense of data likelihood). We are focusing on Maximum Likelihood Linear Transformations and the experiments were performed utilizing the feature Maximum Likelihood Linear Regression (fMLLR). Experiments are aimed to describe the behavior of the system extended by proposed methods.

In this paper we present a novel method for adaptation of a multi-layer perceptron neural network (MLP ANN). Nowadays, the adaptation of the ANN is usually done as an incremental retraining either of a subset or the complete set of the ANN parameters. However, since sometimes the amount of the adaptation data is quite small, there is a fundamental drawback of such approach - during retraining, the network parameters can be easily overfitted to the new data. There certainly are techniques that can help overcome this problem (early-stopping, cross-validation), however application of such techniques leads to more complex and possibly more data hungry training procedure. The proposed method approaches the problem from a different perspective. We use the fact that in many cases we have an additional knowledge about the problem. Such additional knowledge can be used to limit the dimensionality of the adaptation problem. We applied the proposed method on speaker adaptation of a phoneme recognizer based on TRAPS (Temporal Patterns) parameters. We exploited the fact that the employed TRAPS parameters are constructed using log-outputs of mel-filter bank and by virtue of reformulating the first layer weight matrix adaptation problem as a mel-filter bank output adaptation problem, we were able to significantly limit the number of free variables. Adaptation using the proposed method resulted in a substantial improvement of phoneme recognizer accuracy.

This paper deals with a combination of basic adaptation techniques of Hidden Markov Model used in the speech recognition. The adaptation methods approach the data only through their statistics, which have to be accumulated before the adaptation process. When performing two adaptations subsequently, the data statistics have to be accumulated twice in each of the adaptation passes. However, when the adaptation methods are chosen with care, the data statistics may be accumulated only once, as proposed in this paper. This significantly reduces the time consumption and avoids the need to store all the adaptation data. Combination of Maximum A-Posteriori Probability and feature Maximum Likelihood Linear Regression adaptation is considered. Motivation for such an approach could be the on-line adaptation, where the time consumption is of big importance.

This article describes advances in Czech --- Signed Speech translation. A method using a new criterion based on minimal loss principle for log-linear model phrase extraction was introduced and it was evaluated against two another criteria. The performance of phrase table extracted with introduced method was compared with performance of two another phrase tables (manually and automatically extracted). A new criterion for semantic agreement evaluation of translations was introduced too.

In this paper we describe a HMM-based sign language recognition (SLR) system for isolated signs. In the first part we describe the image parametrization method producing features used for recognition. Our goal was to find the best combination of a feature space dimension reduction method and an HMM structure.

Recent research progress in developing of the Czech - Sign Speech synthesizer is presented. The current goal is to improve the system for automatic synthesis to produce accurate synthesis of the Sign Speech. The synthesis system converts written text to an animation of an artificial human model (avatar). This includes translation of text to sign phrases and their conversion to the animation of the avatar. The animation is composed of movements and deformations of segments of hands, a head and also a face. The system has been evaluated by two initial perceptual tests. The perceptual tests indicate that the designed synthesis system is capable to produce the intelligible Sign Speech.

The aim of this paper is to present an extension of the hidden vector state semantic parser. First, we describe the statistical semantic parsing and its decomposition into the semantic and the lexical model. Subsequently, we present the original hidden vector state parser. Then, we modify its lexical model so that it supports the use of the input sequence of feature vectors instead of the sequence of words. We compose the feature vector from the automatically generated linguistic features (lemma form and morphological tag of the original word). We also examine the effect of including the original word into the feature vector. Finally, we evaluate the modified semantic parser on the Czech Human-Human train timetable corpus. We found that the performance of the semantic parser improved significantly compared with the baseline hidden vector state parser.

The paper describes the system built by the team from the University of West Bohemia for participation in the CLEF 2006 CL-SR track. We have decided to concentrate only on the monolingual searching in the Czech test collection and investigate the effect of proper language processing on the retrieval performance. We have employed the Czech morphological analyser and tagger for that purposes. For the actual search system, we have used the classical tf.idf approach with blind relevance feedback as implemented in the Lemur toolkit. The results indicate that a suitable linguistic preprocessing is indeed crucial for the Czech IR performance.

This paper describes a LVCSR system for automatic online subtitling (closed captioning) of TV transmissions of the Czech Parliament meetings The recognition system is based on Hidden Markov Models, lexical trees and bigram language model The acoustic model is trained on 40 hours of parliament speech and the language model on more than 10M tokens of parliament speech trancriptions The first part of the article is focused on text normalization and class-based language model preparation The second part describes the recognition network and its decoding with respect to real-time operation demands using up to 100k vocabulary The third part outlines the application framework allowing generation and displaying of subtitles for any audio/video source Finally, experimental results obtained on parliament speeches with recognition accuracy varying from 80 to 95 % (according to the discussed topic) are reported and discussed.

In the paper we demonstrate a complex supervised learning method based on a binary decision graphs This method is employed in construction of a silence/speech detector Performance of the resulting silence/speech detector is compared with performance of common silence/speech detectors used in telecommunications and with a detector based on HMM and a bigram silence/speech language model Each non-leaf node of a decision graph has assigned a question and a sub-classifier answering this question We test three kinds of these sub-classifiers: linear classifier, classifier based on separating quadratic hyper-plane (SQHP), and Support Vector Machines (SVM) based classifier Moreover, besides usage of a single decision graph we investigate application of a set of binary decision graphs.

In this paper we describe the part of the text preprocessing module in our text-to-speech synthesis system which converts numerals written as figures into a readable full-length form, which could be processed by a phonetic transcription module. The numerals conversion is a significant issue in inflectional language as Czech, Russian or Slovak because morphological and semantic information is necessary to make the conversion unambiguous. In the paper three part-of-speech tagging methods are compared. Furthermore, a method reducing the tagset to increase the numerals conversion accuracy is presented in the paper.

This paper is devoted to letter-to-sound (LTS) conversion for the German language. The system for automatic LTS conversion is an essential module in speech recognition and text-to-speech (TTS) systems. The LTS rules are generated by the decision tree based method. Although this method usually uses the entropy criterion we proposed and used a new original criterion described in the paper. The both criterions are investigated and their results are compared. Our training lexicon does not contain any alignment information. Therefore, three different automatic alignment methods were used and tested. A chunk-based alignment instead of usual one-to-one character correspondence is employed.

This paper deals with a lemmatization technique and its using for phonetic transcription of exceptional words. The lemmatizer is based on language morphology and uses a lexicon of basic word forms and a set of inversion derivation rules to acquire lemmatization rules, which are essential for finding word bases. The lemmatization algorithm and its necessary modifications for transcription of exceptional words are described. The main goal of the designed system is to save computer memory for exceptional lexicon storing. The experimental results showed that it is possible to save from 18.3% (English) to 98.4% (Finnish) of the full lexicon size. Hence, the described technique can be applied with advantage for high inflectional and agglutinative languages.

In this paper a technique for detection and rejection of incorrectly recognized words is described. The speech recognition system we used is based on a speaker-independent continuous density Hidden Markov Model recognizer and so-called mumble model, which structure and function is also depicted. An improved rejection technique is presented in comparison with the heuristic rejection method that we previously used. The new method is fully statistically based. Therefore selection of features for training and classification, procedures for statistical models parameters estimation, and experimental results are reported. The improved rejection technique achieves approximately 12% error rate in detection of incorrectly recognized words.

Speech/non-speech (S/NS) detection plays the important role for automatic speech recognition (ASR) system, especially in the case of isolated words or commands recognition. Even in continuous speech a S/NS decision can be made at the beginning and at the end of a sequence resulting in a "sleep mode" of the speech recognizer during the silence and in a reduction of computation demands. It is very difficult, however, to precisely locate the endpoints of the input utterance because of unpredictable background noise. In the proposed method in this paper, we make use of the advantages of two approaches (i.e. to try to find the best set of heuristic features and apply a statistical induction method) for the best S/NS decision.

Aim: Obesity and physical inactivity had already reached epidemic, becoming one of the leading causes of death and disability worldwide. The objective of this study was to investigate the current level and a nearly three decades’ trend of muscular fitness (MF) and the nutritional state of Czech children. Methods: In 2013, cross-sectional epidemiological survey was conducted in schoolchildren with examination of the anthropometric charac-teristics and measured MF using a UNIFIT test, which normative for MF categories was given as results of national monitoring data collected in 1987. In total 896 current schoolchildren (472 aged 8–9 years and 424 aged 12–13 years) were examined from selected children through stratified sampling from the Czech Republic. Results: Against 31% of children from 1986 there were 74% of current children classified as “poor” or “below normal” in the MF category (p < 0.001). MF was inversely associated with BMI. Poor posture was diagnosed in 24% of children, more frequently in overweight (OW) and obese (OB) children in comparison to normal body weight children (p < 0.001). In comparison to 10% of OW and OB children in 1991, using these cut-off points for body mass index references, the current prevalence of OW and OB doubled (19.8%). Conclusions: Current Czech schoolchildren showed a doubled prevalence of OW and OB during the last two decades and simultaneously during nearly three decades there were more than doubled prevalence of “poor” or “under normal” MF of children, with overall dramatic decrease of MF in current schoolchildren.