Speaker diarization: Segmentation and clustering of speeches

Speaker diarization is the process of sorting speeches according to the speaker. Diarization helps to search and retrieve what a certain speaker uttered in a meeting. Applications of diarization systemsextend to other domains than meetings, for example, lectures, telephone, television, and radio. Besides, diarization enhances the performance of several speech technologies such as speaker recognition, automatic transcription, and speaker tracking. Methodologies previously used in developing diarization systems are discussed. Prior results and techniques are studied and compared. Methods such as Hidden Markov Models and Gaussian Mixture Models that are used in speaker recognition and other speech technologies are also used in speaker diarization. The objective of this thesis is to develop a speaker diarization system in meeting domain. Experimental part of this work indicates that zero-crossing rate can be used effectively in breaking down the audio stream into segments, and adaptive Gaussian Models fit adequately short audio segments. Results show that 35 Gaussian Models and one second as average length of each segment are optimum values to build a diarization system for the tested data. Uniting the segments which are uttered by same speaker is done in a bottom-up clustering by a newapproach of categorizing the mixture weights.

Comparison of Local Feature Detectors and Descriptors for Visual Object Categorization

Local features are used in many computer vision tasks including visual object categorization, content-based image retrieval and object recognition to mention a few. Local features are points, blobs or regions in images that are extracted using a local feature detector. To make use of extracted local features the localized interest points are described using a local feature descriptor. A descriptor histogram vector is a compact representation of an image and can be used for searching and matching images in databases. In this thesis the performance of local feature detectors and descriptors is evaluated for object class detection task. Features are extracted from image samples belonging to several object classes. Matching features are then searched using random image pairs of a same class. The goal of this thesis is to find out what are the best detector and descriptor methods for such task in terms of detector repeatability and descriptor matching rate.

Kahden merkkijonon pisimmän yhteisen alijonon ongelma ja sen ratkaiseminen

Tämä tutkielma kuuluu merkkijonoalgoritmiikan piiriin. Merkkijono S on merkkijonojen X[1..m] ja Y[1..n] yhteinen alijono, mikäli se voidaan muodostaa poistamalla X:stä 0..m ja Y:stä 0..n kappaletta merkkejä mielivaltaisista paikoista. Jos yksikään X:n ja Y:n yhteinen alijono ei ole S:ää pidempi, sanotaan, että S on X:n ja Y:n pisin yhteinen alijono (lyh. PYA). Tässä työssä keskitytään kahden merkkijonon PYAn ratkaisemiseen, mutta ongelma on yleistettävissä myös useammalle jonolle. PYA-ongelmalle on sovelluskohteita – paitsi tietojenkäsittelytieteen niin myös bioinformatiikan osa-alueilla. Tunnetuimpia niistä ovat tekstin ja kuvien tiivistäminen, tiedostojen versionhallinta, hahmontunnistus sekä DNA- ja proteiiniketjujen rakennetta vertaileva tutkimus. Ongelman ratkaisemisen tekee hankalaksi ratkaisualgoritmien riippuvuus syötejonojen useista eri parametreista. Näitä ovat syötejonojen pituuden lisäksi mm. syöttöaakkoston koko, syötteiden merkkijakauma, PYAn suhteellinen osuus lyhyemmän syötejonon pituudesta ja täsmäävien merkkiparien lukumäärä. Täten on vaikeaa kehittää algoritmia, joka toimisi tehokkaasti kaikille ongelman esiintymille. Tutkielman on määrä toimia yhtäältä käsikirjana, jossa esitellään ongelman peruskäsitteiden kuvauksen jälkeen jo aikaisemmin kehitettyjä tarkkoja PYAalgoritmeja. Niiden tarkastelu on ryhmitelty algoritmin toimintamallin mukaan joko rivi, korkeuskäyrä tai diagonaali kerrallaan sekä monisuuntaisesti prosessoiviin. Tarkkojen menetelmien lisäksi esitellään PYAn pituuden ylä- tai alarajan laskevia heuristisia menetelmiä, joiden laskemia tuloksia voidaan hyödyntää joko sellaisinaan tai ohjaamaan tarkan algoritmin suoritusta. Tämä osuus perustuu tutkimusryhmämme julkaisemiin artikkeleihin. Niissä käsitellään ensimmäistä kertaa heuristiikoilla tehostettuja tarkkoja menetelmiä. Toisaalta työ sisältää laajahkon empiirisen tutkimusosuuden, jonka tavoitteena on ollut tehostaa olemassa olevien tarkkojen algoritmien ajoaikaa ja muistinkäyttöä. Kyseiseen tavoitteeseen on pyritty ohjelmointiteknisesti esittelemällä algoritmien toimintamallia hyvin tukevia tietorakenteita ja rajoittamalla algoritmien suorittamaa tuloksetonta laskentaa parantamalla niiden kykyä havainnoida suorituksen aikana saavutettuja välituloksia ja hyödyntää niitä. Tutkielman johtopäätöksinä voidaan yleisesti todeta tarkkojen PYA-algoritmien heuristisen esiprosessoinnin lähes systemaattisesti pienentävän niiden suoritusaikaa ja erityisesti muistintarvetta. Lisäksi algoritmin käyttämällä tietorakenteella on ratkaiseva vaikutus laskennan tehokkuuteen: mitä paikallisempia haku- ja päivitysoperaatiot ovat, sitä tehokkaampaa algoritmin suorittama laskenta on.

Real-time Image-based RGB-D Camera Motion Tracking and Environment Mapping

In this work, image based estimation methods, also known as direct methods, are studied which avoid feature extraction and matching completely. Cost functions use raw pixels as measurements and the goal is to produce precise 3D pose and structure estimates. The cost functions presented minimize the sensor error, because measurements are not transformed or modified. In photometric camera pose estimation, 3D rotation and translation parameters are estimated by minimizing a sequence of image based cost functions, which are non-linear due to perspective projection and lens distortion. In image based structure refinement, on the other hand, 3D structure is refined using a number of additional views and an image based cost metric. Image based estimation methods are particularly useful in conditions where the Lambertian assumption holds, and the 3D points have constant color despite viewing angle. The goal is to improve image based estimation methods, and to produce computationally efficient methods which can be accomodated into real-time applications. The developed image-based 3D pose and structure estimation methods are finally demonstrated in practise in indoor 3D reconstruction use, and in a live augmented reality application.

Backpacker 2.0 - Contemporary backpackers and their use of social media as an information and communication platform

Long-term independent budget travel to countries far away has become increasingly common over the last few decades, and backpacking has now entered the tourism mainstream. Nowadays, backpackers are a very important segment of the global travel market. Backpacking is a type of tourism that involves a lot of information search activities. The Internet has become a major source of information as well as a platform for tourism business transactions. It allows travelers to gain information very effortlessly and to learn about tourist destinations and products directly from other travelers in the form of electronic word-of-mouth (eWOM). Social media has penetrated and changed the backpacker market, as now modern travelers can stay connected to people at home, read online recommendations, and organize and book their trips very independently. In order to create a wider understanding on modern-day backpackers and their information search and share behavior in the Web 2.0 era, this thesis examined contemporary backpackers and their use of social media as an information and communication platform. In order to achieve this goal, three sub-objectives were identified: 1. to describe contemporary backpacker tourism 2. to examine contemporary backpackers’ travel information search and share behavior 3. to explore the impacts of new information and communications technologies and Web 2.0 on backpacker tourism The empirical data was gathered with an online survey, thus the method of analysis was mainly quantitative, and a qualitative method was used for a brief analysis of open questions. The research included both descriptive and analytical approaches, as the goal was to describe modern-day backpackers, and to examine possible interdependencies between information search and share behavior and background variables. The interdependencies were tested for statistical significance with the help of five research hypotheses. The results suggested that backpackers no longer fall under the original backpacker definitions described some decades ago. Now, they are mainly short-term travelers, whose trips resemble more those of mainstream tourists. They use communication technologies very actively, and particularly social media. Traditional information sources, mainly guide books and recommendations from friends, are of great importance to them but also eWOM sources are widely used in travel decision making. The use of each source varies according to the stage of the trip. All in all, Web 2.0 and new ICTs have transformed the backpacker tourism industry in many ways. Although the experience has become less authentic in some travelers’ eyes, the backpacker culture is still recognizable.

Self-deposit, discovery, and delivery of scientific GIS datasets using GeoHydra

Presentation at Open Repositories 2014, Helsinki, Finland, June 9-13, 2014

Search engine advertising in web retailing : an efficiency analysis

This study examines the efficiency of search engine advertising strategies employed by firms. The research setting is the online retailing industry, which is characterized by extensive use of Web technologies and high competition for market share and profitability. For Internet retailers, search engines are increasingly serving as an information gateway for many decision-making tasks. In particular, Search engine advertising (SEA) has opened a new marketing channel for retailers to attract new customers and improve their performance. In addition to natural (organic) search marketing strategies, search engine advertisers compete for top advertisement slots provided by search brokers such as Google and Yahoo! through keyword auctions. The rationale being that greater visibility on a search engine during a keyword search will capture customers' interest in a business and its product or service offerings. Search engines account for most online activities today. Compared with the slow growth of traditional marketing channels, online search volumes continue to grow at a steady rate. According to the Search Engine Marketing Professional Organization, spending on search engine marketing by North American firms in 2008 was estimated at $13.5 billion. Despite the significant role SEA plays in Web retailing, scholarly research on the topic is limited. Prior studies in SEA have focused on search engine auction mechanism design. In contrast, research on the business value of SEA has been limited by the lack of empirical data on search advertising practices. Recent advances in search and retail technologies have created datarich environments that enable new research opportunities at the interface of marketing and information technology. This research uses extensive data from Web retailing and Google-based search advertising and evaluates Web retailers' use of resources, search advertising techniques, and other relevant factors that contribute to business performance across different metrics. The methods used include Data Envelopment Analysis (DEA), data mining, and multivariate statistics. This research contributes to empirical research by analyzing several Web retail firms in different industry sectors and product categories. One of the key findings is that the dynamics of sponsored search advertising vary between multi-channel and Web-only retailers. While the key performance metrics for multi-channel retailers include measures such as online sales, conversion rate (CR), c1ick-through-rate (CTR), and impressions, the key performance metrics for Web-only retailers focus on organic and sponsored ad ranks. These results provide a useful contribution to our organizational level understanding of search engine advertising strategies, both for multi-channel and Web-only retailers. These results also contribute to current knowledge in technology-driven marketing strategies and provide managers with a better understanding of sponsored search advertising and its impact on various performance metrics in Web retailing.

Retrial Queues with Orbital Search

Queueing system in which arriving customers who find all servers and waiting positions (if any) occupied many retry for service after a period of time are retrial queues or queues with repeated attempts. This study deals with two objectives one is to introduce orbital search in retrial queueing models which allows to minimize the idle time of the server. If the holding costs and cost of using the search of customers will be introduced, the results we obtained can be used for the optimal tuning of the parameters of the search mechanism. The second one is to provide insight of the link between the corresponding retrial queue and the classical queue. At the end we observe that when the search probability Pj = 1 for all j, the model reduces to the classical queue and when Pj = 0 for all j, the model becomes the retrial queue. It discusses the performance evaluation of single-server retrial queue. It was determined by using Poisson process. Then it discuss the structure of the busy period and its analysis interms of Laplace transforms and also provides a direct method of evaluation for the first and second moments of the busy period. Then it discusses the M/ PH/1 retrial queue with disaster to the unit in service and orbital search, and a multi-server retrial queueing model (MAP/M/c) with search of customers from the orbit. MAP is convenient tool to model both renewal and non-renewal arrivals. Finally the present model deals with back and forth movement between classical queue and retrial queue. In this model when orbit size increases, retrial rate also correspondingly increases thereby reducing the idle time of the server between services

Development of a Biometric Personal Authentication System Based on Fingerprint and Speech

Biometrics deals with the physiological and behavioral characteristics of an individual to establish identity. Fingerprint based authentication is the most advanced biometric authentication technology. The minutiae based fingerprint identification method offer reasonable identification rate. The feature minutiae map consists of about 70-100 minutia points and matching accuracy is dropping down while the size of database is growing up. Hence it is inevitable to make the size of the fingerprint feature code to be as smaller as possible so that identification may be much easier. In this research, a novel global singularity based fingerprint representation is proposed. Fingerprint baseline, which is the line between distal and intermediate phalangeal joint line in the fingerprint, is taken as the reference line. A polygon is formed with the singularities and the fingerprint baseline. The feature vectors are the polygonal angle, sides, area, type and the ridge counts in between the singularities. 100% recognition rate is achieved in this method. The method is compared with the conventional minutiae based recognition method in terms of computation time, receiver operator characteristics (ROC) and the feature vector length. Speech is a behavioural biometric modality and can be used for identification of a speaker. In this work, MFCC of text dependant speeches are computed and clustered using k-means algorithm. A backpropagation based Artificial Neural Network is trained to identify the clustered speech code. The performance of the neural network classifier is compared with the VQ based Euclidean minimum classifier. Biometric systems that use a single modality are usually affected by problems like noisy sensor data, non-universality and/or lack of distinctiveness of the biometric trait, unacceptable error rates, and spoof attacks. Multifinger feature level fusion based fingerprint recognition is developed and the performances are measured in terms of the ROC curve. Score level fusion of fingerprint and speech based recognition system is done and 100% accuracy is achieved for a considerable range of matching threshold

Uncertainty Propagation in Model-Based Recognition

Building robust recognition systems requires a careful understanding of the effects of error in sensed features. Error in these image features results in a region of uncertainty in the possible image location of each additional model feature. We present an accurate, analytic approximation for this uncertainty region when model poses are based on matching three image and model points, for both Gaussian and bounded error in the detection of image points, and for both scaled-orthographic and perspective projection models. This result applies to objects that are fully three- dimensional, where past results considered only two-dimensional objects. Further, we introduce a linear programming algorithm to compute the uncertainty region when poses are based on any number of initial matches. Finally, we use these results to extend, from two-dimensional to three- dimensional objects, robust implementations of alignmentt interpretation- tree search, and ransformation clustering.

COMP6049 Surveys Lecture and Tutorial

set of slides and matching notes for lecture and subsequent linked tutorial introducing an overview of designing and implementing surveys

''Rainbow Designer'' : for global and multicultural design

What do the designers tend to achieve? To relate themselves to the reality by producing visual registers of emotions and thoughts, or by projecting and producing objects that are functional, adapting technologies to daily needs. That requires that a designer be a keen observer of his physical surroundings and have a fine sensibility to cultures, enabling him to disassemble the latent forms of the reality and cultural symbolisms in order to perceive the order underlying them and the principles of their composition and unity. Only then could he reproduce the nature and respond to cultural callings. In this process of understanding the surrounding reality of nature and cultures, a designer always moves, generally without being aware of it, between two processes: identity search and self-identification.

Molecular dynamics simulation of interactions between a sodium dodecyl sulfate micelle and a poly(ethylene oxide) polymer

We have performed atomistic molecular dynamics simulations of an anionic sodium dodecyl sulfate (SDS) micelle and a nonionic poly(ethylene oxide) (PEO) polymer in aqueous solution. The micelle consisted of 60 surfactant molecules, and the polymer chain lengths varied from 20 to 40 monomers. The force field parameters for PEO were adjusted by using 1,2-dimethoxymethane (DME) as a model compound and matching its hydration enthalpy and conformational behavior to experiment. Excellent agreement with previous experimental and simulation work was obtained through these modifications. The simulated scaling behavior of the PEO radius of gyration was also in close agreement with experimental results. The SDS-PEO simulations show that the polymer resides on the micelle surface and at the hydrocarbon-water interface, leading to a selective reduction in the hydrophobic contribution to the solvent-accessible surface area of the micelle. The association is mainly driven by hydrophobic interactions between the polymer and surfactant tails, while the interaction between the polymer and sulfate headgroups on the micelle surface is weak. The 40-monomer chain is mostly wrapped around the micelle, and nearly 90% of the monomers are adsorbed at low PEO concentration. Simulations were also performed with multiple 20-monomer chains, and gradual addition of polymer indicates that about 120 monomers are required to saturate the micelle surface. The stoichiometry of the resulting complex is in close agreement with experimental results, and the commonly accepted "beaded necklace" structure of the SDS-PEO complex is recovered by our simulations.

Flexible MMSE DFE using length and LAG adaptation

This paper investigates how to choose the optimum tap-length and decision delay for the decision feedback equalizer (DFE). Although the feedback filter length can be set as the channel memory, there is no closed-form expression for the feedforward filter length and decision delay. In this paper, first we analytically show that the two dimensional search for the optimum feedforward filter length and decision delay can be simplified to a one dimensional search, and then describe a new adaptive DFE where the optimum structural parameters can be self-adapted.