Dynamic Time Warping (DTW) é uma técnica do tipo pattern matching para reconhecimento de padrões de voz, sendo baseada no alinhamento temporal de um sinal com os diversos modelos de referência. Uma desvantagem da DTW é o seu alto custo computacional. Este trabalho apresenta uma versão da DTW que, utilizando a Transformada Wavelet Discreta (DWT), reduz a sua complexidade. O desempenho obtido com a proposta foi muito promissor, ganhando em termos de velocidade de reconhecimento e recursos de memória consumidos, enquanto a precisão da DTW não é afetada. Os testes foram realizados com alguns fonemas extraídos da base de dados TIMIT do Linguistic Data Consortium (LDC) / Dynamic TimeWarping (DTW) is a pattern matching technique for speech recognition, that is based on a temporal alignment of the input signal with the template models. One drawback of this technique is its high computational cost. This work presents a modified version of the DTW, based on the DiscreteWavelet Transform (DWT), that reduces the complexity of the original algorithm. The performance obtained with the proposed algorithm is very promising, improving the recognition in terms of time and memory allocation, while the precision is not affected. Tests were performed with speech data collected from TIMIT corpus provided by Linguistic Data Consortium (LDC).

Les séries temporelles d’images satellites (STIS) sont des données cruciales pour l’observation de la terre. Les séries temporelles actuelles sont soit des séries à haute résolution temporelle (Spot-Végétation, MODIS), soit des séries à haute résolution spatiale (Landsat). Dans les années à venir, les séries temporelles d’images satellites à hautes résolutions spatiale et temporelle vont être produites par le programme Sentinel de l’ESA. Afin de traiter efficacement ces immenses quantités de données qui vont être produites (par exemple, Sentinel-2 couvrira la surface de la terre tous les cinq jours, avec des résolutions spatiales allant de 10m à 60m et disposera de 13 bandes spectrales), de nouvelles méthodes ont besoin d’être développées. Cette thèse se focalise sur la comparaison des profils d’évolution radiométrique, et plus précisément la mesure de similarité « Dynamic Time Warping », qui constitue un outil permettant d’exploiter la structuration temporelle des séries d’images satellites. / Satellite Image Time Series are becoming increasingly available and will continue to do so in the coming years thanks to the launch of space missions, which aim at providing a coverage of the Earth every few days with high spatial resolution (ESA’s Sentinel program). In the case of optical imagery, it will be possible to produce land use and cover change maps with detailed nomenclatures. However, due to meteorological phenomena, such as clouds, these time series will become irregular in terms of temporal sampling. In order to consistently handle the huge amount of information that will be produced (for instance, Sentinel-2 will cover the entire Earth’s surface every five days, with 10m to 60m spatial resolution and 13 spectral bands), new methods have to be developed. This Ph.D. thesis focuses on the “Dynamic Time Warping” similarity measure, which is able to take the most of the temporal structure of the data, in order to provide an efficient and relevant analysis of the remotely observed phenomena.

Microsoft Kinect įrenginys išleistas tik 2010 metais. Jis buvo skirtas Microsoft Xbox 360 vaizdo žaidimų konsolei, vėliau 2012 metais buvo pristatytas Kinect ir Windows personaliniams kompiuteriams. Taigi tai palyginus naujas įrenginys ir aktualus šiai dienai. Daugiausiai yra sukurta kompiuterinių žaidimų, kurie naudoja Microsoft Kinect įrenginį, bet šį įrenginį galima panaudoti daug plačiau ne tik žaidimuose, viena iš sričių tai sportas, konkrečiau treniruotės, kurias būtų galima atlikti namuose. Šiuo metu pasaulyje yra programinės įrangos, žaidimų, sportavimo programų, kuri leidžia kontroliuoti treniruočių eigą sekdama ar žmogus teisingai atlieka treniruotėms numatytus judesius. Kadangi Lietuvoje panašios programinės įrangos nėra, taigi reikia sukurti įrangą, kuri leistų Lietuvos treneriams kurti treniruotes orientuotas į šio įrenginio panaudojimą. Šio darbo pagrindinis tikslas yra atlikti Kinect įrenginiui skirtų gestų atpažinimo algoritmų tyrimą, kaip tiksliai jie gali atpažinti gestus ar gestą. Pagrindinis dėmesys skiriamas šiai problemai, taip pat keliami, bet netyrinėjami kriterijai kaip atpažinimo laikas, bei realizacijos sunkumas. Šiame darbe sukurta programa, judesius bei gestus atpažįsta naudojant Golden Section Search algoritmą. Algoritmas palygina du modelius ar šablonus, ir jei neranda atitikmens, tai pirmasis šablonas šiek tiek pasukamas ir lyginimo procesas paleidžiamas vėl, taipogi tam tikro kintamojo dėka galime keisti algoritmo tikslumą. Taipogi... [toliau žr. visą tekstą] / Microsoft Kinect device was released in 2010. It was designed for Microsoft Xbox 360 gaming console, later on in 2012 was presented Kinect device for Windows personal computer. So this device is new and current. Many games has been created for Microsoft Kinect device, but this device could be used not only in games, one of the areas where we can use it its sport, specific training, which can be performed at home. At this moment in world are huge variety of games, software, training programs which allows user to control training course by following a person properly perform training provided movements. Since in Lithuania similar software is not available, so it is necessary to create software that would allow Lithuania coaches create training focused on the use of this device. The main goal of this work is to perform research of the Kinect device gesture recognition algorithms to study exactly how they can recognize gestures or gesture. It will focus on this issue mainly, but does not address the criteria for recognition as the time and difficulty of realization. In this paper, a program that recognizes movements and gestures are using the Golden section search algorithm. Algorhithm compares the two models or templates, and if it can not find a match, this is the first template slightly rotated and comparison process is started again, also a certain variable helping, we can modify the algorithm accuracy. Also for comparison we can use Hidden Markov models algorhithm received... [to full text]

Submitted by Jordan (jordanbiblio@gmail.com) on 2017-05-04T13:38:27Z No. of bitstreams: 1 DISS_2015_Gracyeli Santos Souza Guarienti.pdf: 4160382 bytes, checksum: 066e507b4df1c012a091983043416a9b (MD5) / Approved for entry into archive by Jordan (jordanbiblio@gmail.com) on 2017-05-04T15:41:01Z (GMT) No. of bitstreams: 1 DISS_2015_Gracyeli Santos Souza Guarienti.pdf: 4160382 bytes, checksum: 066e507b4df1c012a091983043416a9b (MD5) / Made available in DSpace on 2017-05-04T15:41:01Z (GMT). No. of bitstreams: 1 DISS_2015_Gracyeli Santos Souza Guarienti.pdf: 4160382 bytes, checksum: 066e507b4df1c012a091983043416a9b (MD5) Previous issue date: 2015-05-27 / CAPES / Variáveis climatológicas podem ser estudadas a partir de seu comportamento temporal. Nesse sentido, este trabalho desenvolveu uma técnica computacional de processamento espaço-temporal de variáveis climatológicas que utiliza busca por similaridade e a possibilidade de comparação em várias resoluções temporais. Para demonstração do uso da técnica e verificação dos resultados, sequências de processamento foram aplicadas em séries de precipitação de um período de quinze anos usando os algoritmos Dynamic Time Warping (DTW) e wavelet em quatro biomas: Amazônia, Cerrado, Pantanal e Mata Atlântica. A técnica foi aplicada nas séries originais e em suas wavelets, com resoluções temporais mensal, semestral, anual e quinze anos de forma a permitir que análises específicas em cada resolução possam ser aplicadas. A flexibilidade e a variedade de resoluções temporais permitidas pela técnica torna possível acrescentar aos processos de monitoramento ambiental novas perspectivas em tomadas de decisão. / Climatic variables can be studied from its temporal behavior. In this sense, this study developed a temporal analysis technique for climatological variables using similarity search and the possibility of comparison in various temporal resolution levels. For the income statement, several processing sequences were applied in series of precipitation a period of fifteen years using the Dynamic Time Warping algorithm (DTW) and wavelet on four biomes: Amazon, Cerrado, Pantanal and Atlantic Forest. The technique was applied to the original data and wavelets, in the temporal resolution of time monthly, semi-annual, annual and fifteen years enable visualization and comparison of data on these different scales. Application the technique developed in this study, provide new perspectives to decision-making in environmental monitoring processes.

Dynamic Time Warping (DTW) é uma técnica do tipo pattern matching para reconhecimento de padrões de voz, sendo baseada no alinhamento temporal de um sinal com os diversos modelos de referência. Uma desvantagem da DTW é o seu alto custo computacional. Este trabalho apresenta uma versão da DTW que, utilizando a Transformada Wavelet Discreta (DWT), reduz a sua complexidade. O desempenho obtido com a proposta foi muito promissor, ganhando em termos de velocidade de reconhecimento e recursos de memória consumidos, enquanto a precisão da DTW não é afetada. Os testes foram realizados com alguns fonemas extraídos da base de dados TIMIT do Linguistic Data Consortium (LDC) / Dynamic TimeWarping (DTW) is a pattern matching technique for speech recognition, that is based on a temporal alignment of the input signal with the template models. One drawback of this technique is its high computational cost. This work presents a modified version of the DTW, based on the DiscreteWavelet Transform (DWT), that reduces the complexity of the original algorithm. The performance obtained with the proposed algorithm is very promising, improving the recognition in terms of time and memory allocation, while the precision is not affected. Tests were performed with speech data collected from TIMIT corpus provided by Linguistic Data Consortium (LDC).

Le suivi de partition est défini comme étant la synchronisation sur ordinateur entre une partition musicale connue et le signal sonore de l'interprète de cette partition. Dans le cas particulier de la voix chantée, il y a encore place à l'amélioration des algorithmes existants, surtout pour le suivi de partition en temps réel. L'objectif de ce projet est donc d'arriver à mettre en oeuvre un logiciel suiveur de partition robuste et en temps-réel utilisant le signal numérisé de voix chantée et le texte des chansons. Le logiciel proposé utilise à la fois plusieurs caractéristiques de la voix chantée (énergie, correspondance avec les voyelles et nombre de passages par zéro du signal) et les met en correspondance avec la partition musicale en format MusicXML. Ces caractéristiques, extraites pour chaque trame, sont alignées aux unités phonétiques de la partition. En parallèle avec cet alignement à court terme, le système ajoute un deuxième niveau d'estimation plus fiable sur la position en associant une segmentation du signal en blocs de chant à des sections chantées en continu dans la partition. La performance du système est évaluée en présentant les alignements obtenus en différé sur 3 extraits de chansons interprétés par 2 personnes différentes, un homme et une femme, en anglais et en français.

Given the ubiquity of time series data, and the exponential growth of databases, there has recently been an explosion of interest in time series data mining. Finding similar trends and patterns among time series data is critical for many applications ranging from financial planning, weather forecasting, stock analysis to policy making. With time series being high-dimensional objects, detection of similar trends especially at the granularity of subsequences or among time series of different lengths and temporal misalignments incurs prohibitively high computation costs. Finding trends using non-metric correlation measures further compounds the complexity, as traditional pruning techniques cannot be directly applied. My dissertation addresses these challenges while meeting the need to achieve near real-time responsiveness. First, for retrieving exact similarity results using Lp-norm distances, we design a two-layered time series index for subsequence matching. Time series relationships are compactly organized in a directed acyclic graph embedded with similarity vectors capturing subsequence similarities. Powerful pruning strategies leveraging the graph structure greatly reduce the number of time series as well as subsequence comparisons, resulting in a several order of magnitude speed-up. Second, to support a rich diversity of correlation analytics operations, we compress time series into Euclidean-based clusters augmented by a compact overlay graph encoding correlation relationships. Such a framework supports a rich variety of operations including retrieving positive or negative correlations, self correlations and finding groups of correlated sequences. Third, to support flexible similarity specification using computationally expensive warped distance like Dynamic Time Warping we design data reduction strategies leveraging the inexpensive Euclidean distance with subsequent time warped matching on the reduced data. This facilitates the comparison of sequences of different lengths and with flexible alignment still within a few seconds of response time. Comprehensive experimental studies using real-world and synthetic datasets demonstrate the efficiency, effectiveness and quality of the results achieved by our proposed techniques as compared to the state-of-the-art methods.

The analysis of sequential data is important in business, science, and engineering, for tasks such as signal processing, user behavior mining, and commercial transactions analysis. In this dissertation, we build upon the Collective Dynamical Modeling and Clustering (CDMC) framework for discrete time series modeling, by making contributions to clustering initialization, dynamical modeling, and scaling. We first propose a modified Dynamic Time Warping (DTW) approach for clustering initialization within CDMC. The proposed approach provides DTW metrics that penalize deviations of the warping path from the path of constant slope. This reduces over-warping, while retaining the efficiency advantages of global constraint approaches, and without relying on domain dependent constraints. Second, we investigate the use of semi-Markov chains as dynamical models of temporal sequences in which state changes occur infrequently. Semi-Markov chains allow explicitly specifying the distribution of state visit durations. This makes them superior to traditional Markov chains, which implicitly assume an exponential state duration distribution. Third, we consider convergence properties of the CDMC framework. We establish convergence by viewing CDMC from an Expectation Maximization (EM) perspective. We investigate the effect on the time to convergence of our efficient DTW-based initialization technique and selected dynamical models. We also explore the convergence implications of various stopping criteria. Fourth, we consider scaling up CDMC to process big data, using Storm, an open source distributed real-time computation system that supports batch and distributed data processing. We performed experimental evaluation on human sleep data and on user web navigation data. Our results demonstrate the superiority of the strategies introduced in this dissertation over state-of-the-art techniques in terms of modeling quality and efficiency.

Gesture Recognition provides an efficient human-computer interaction for interactive and intelligent computing. In this work, we address the problem of gesture recognition using the theory of random projection and by formulating the recognition problem as an $\ell_1$-minimization problem. The gesture recognition uses a single 3-axis accelerometer for data acquisition and comprises two main stages: a training stage and a testing stage. For training, the system employs dynamic time warping as well as affinity propagation to create exemplars for each gesture while for testing, the system projects all candidate traces and also the unknown trace onto the same lower dimensional subspace for recognition. A dictionary of 18 gestures is defined and a database of over 3,700 traces is created from 7 subjects on which the system is tested and evaluated. Simulation results reveal a superior performance, in terms of accuracy and computational complexity, compared to other systems in the literature.

Gesture Recognition provides an efficient human-computer interaction for interactive and intelligent computing. In this work, we address the problem of gesture recognition using the theory of random projection and by formulating the recognition problem as an $\ell_1$-minimization problem. The gesture recognition uses a single 3-axis accelerometer for data acquisition and comprises two main stages: a training stage and a testing stage. For training, the system employs dynamic time warping as well as affinity propagation to create exemplars for each gesture while for testing, the system projects all candidate traces and also the unknown trace onto the same lower dimensional subspace for recognition. A dictionary of 18 gestures is defined and a database of over 3,700 traces is created from 7 subjects on which the system is tested and evaluated. Simulation results reveal a superior performance, in terms of accuracy and computational complexity, compared to other systems in the literature.