9 research outputs found
Real-time segmentation of 3D echocardiograms using a state estimation approach with deformable models
No full textCardiac ultrasound, or echocardiography, is considered to be the quickest and most cost effective imaging modality for assessment of cardiac function. The modality is unique in that it allows for real-time imaging of the heart, using portable equipment. The latest generation of echocardiography scanners are capable of acquiring dense image volumes in either real time, or over a few heart beats. Methods for analyzing these images, however, are lagging behind, with all existing methods for segmentation of 3D ultrasound data considered too computationally intensive to operate at acquisition rate. Availability of methods for analysis of 3D ultrasound in real-time could open up possibilities for very quick and simple measurement of cardiac function, potentially conducted during image acquisition. Results from automatic methods would also lessen the amount of inter- or intra-examiner differences introduced during the analysis, which in turn could lead to more reproducible results. The left ventricle of the heart has traditionally been the chamber of most interest within the field cardiac image analysis. This thesis has therefore focused on the left ventricle, and investigated ways to measure different aspects of the chamber, such as chamber volumes, myocardial strain and myocardial mass, by means of state estimation techniques.The main goal of this thesis has been to explore the possibilities of using state estimation methods for segmentation and tracking of structures in volumetric data with deformable models. The ability of using non-iterative estimators, such as the Kalman filter, for fitting deformable models to image structures would radically reduce the computationally effort required for performing a 3D segmentation, and open up for real-time usage. An existing Kalman tracking framework have therefore been extended to operate in volumetric data. The framework has been successfully demonstrated to fit ellipsoids, spline surfaces, active-shape surfaces and subdivision surfaces to image data. Furthermore, theory for the simultaneous tracking of several models have been developed. Finally, methods for combining edge-detection measurements with speckle-tracking measurements have been shown, with the potential advantages of material tracking with the lack of drift in edge detection.The framework has been demonstrated to successfully conduct high quality segmentation and tracking of the left ventricle, while operating in real-time without user intervention. Processing times are in the range of milliseconds per frame on standard computer hardware, which is orders of magnitude faster than the state-of-the-art methods. Applications of the framework on 3D echocardiograms has been investigated to prove the feasibility of the framework for automatic analysis of various aspects of cardiac function, such as view alignment, measurement of chamber volumes, myocardial muscle volume and regional myocardial strain. The results for view alignment are considered state of the art, and the results for volume measurements are close to the results of similar studies with semi-automatic tools.PhD i informasjons- og kommunikasjonsteknologiPhD in Information and Communications Technolog
Natural disasters : demographic and health outcomes for women and children
No full textDefence date: 2 February 2023Examining Board: Prof. Juho Härkönen, (EUI, Supervisor); Prof. Fabrizio Bernardi, (Universidad Nacional de Educación a Distancia, former EUI); Assoc. Prof. Kathryn Grace, (University of Minnesota); Assoc. Prof. Tiziana Leone, (The London School of Economics and Political Science)Natural disasters have increased in frequency, intensity and magnitude in recent years, and with climate change more extreme weather events are expected in the future. The aim of this dissertation is to better understand how populations are affected by such disasters by analysing demographic and health outcomes for women and children in low- and lower-middle-income countries. In the first empirical chapter, I and a co-author investigate the impact of the 2010 Haiti earthquake on parity specific fertility by applying event history analysis in a differencein-differences design. The results show no impact from the earthquake on first, second, or third births. In the second empirical chapter, set in the same context, I investigate child nutrition by applying linear regression in a cohort-based difference-in-differences design. The findings show a short-term improvement in child nutrition in less affected areas, and a long-term lack of improvement in child nutrition in the most affected area. In the third empirical chapter, I investigate storms and infant mortality in the Philippines. I apply linear probability models and mother fixed-effects to explore whether storm exposure two months after birth to five years prior to birth has an impact on infant mortality. Overall, the results indicate limited or no association between storms and infant mortality, in both the short- and the long-term. In the final empirical chapter, I and co-authors investigate the associations between floods and maternal healthcare utilisation in Bangladesh. Flood exposure from in utero to five years prior to birth is analysed with a multilevel design, and shows no impact of flood exposure on maternal healthcare utilisation. Findings from this dissertation contribute to a better understanding of the potential consequences of natural disasters for women and children in low- and lower-middleincome countries, which is needed to ensure better targeted measures to protect vulnerable population sub-groups in natural hazard related disasters
The Norwegian Home-building Industry - Locally Embedded or in the Space of Flows?
No full textDuring the last 25 years the neoliberal institutional turn and globalization have caused profound structural changes and challenged our theoretical understandings. In this article the emerging structure of the Norwegian home-building industry is presented, with a focus on the interaction between different actors in the different stages and functions of the home-building network. The structure of employer-contractor relations in the liberalized housing market is disclosed and used as a basis for discussing the theoretical perspectives developed by Amin and Thrift's work on circulation and distanciation and by Castells' work on the network enterprise and space of flows. The questions asked are whether these trends are seen in the Norwegian home-building industry. From the analysis we learn that home-building is considered a local business due to: a capacity restraint regarding local market knowledge; the interaction with local planning authorities; face-to-face meetings; and social relations. On the other hand, there are a set of factors that tell a 'circulation' story: an increasing number of participants, especially in design and planning; a more complex market structure; and a need for developer's knowledge on trends in design and construction. In sum, the analysis discloses a structure of increasing circulation but without a corresponding distanciation, and from Castells' perspective there is an emerging space of flows, but still a rather local geometry; a pattern probably prone to change in the years to come. Copyright (c) 2006 The Author. Journal Compilation (c) 2006 Joint Editors and Blackwell Publishing Ltd.
P2B-14 Real-Time Indication of Acoustic Window for Phased-Array Transducers in Ultrasound Imaging
No full text
Quantitative Cardiac Analysis Algorithms for Pocket-sized Ultrasound Devices
No full textSiden introduksjonen i 1970-årene har ultralyd blitt ett av de viktigste verktøyene for påvisning av hjertesykdommer. De fleste ultralydundersøkelser er hurtige og smertefrie. Ultralydapparatene har tradisjonelt vært store og komplekse systemer, og det har vært nødvendig med omfattende opplæring og erfaring for ˚a anvende dem. Fra årtusenskiftet har det blitt utviklet mindre og mer tilgjengelige ultralydapparater. Dette har resultert i lansering av flere håndholdte systemer. Disse apparatene har et enkelt brukergrensesnitt og relativt lav pris. En forventer at de ogs˚a vil bli brukt av klinikere med mindre ultralyderfaring enn det som har vært tilfelle med de store skannerne.
Parallelt med utviklingen av h˚andholdte systemer har det vært gjort store fremskritt på sanntids 3D-avbildning med ultralyd. Mye av teknologien som har inng˚att i de nye 3Dsystemene, har ogs˚a spilt en nøkkelrolle i miniatyriseringen til h˚andholdte systemer. For sanntids 3D bildesegmentering har det blitt utviklet en svært effektiv algoritme basert på deformerbare modeller og Kalmanfilter. Denne løsningen burde være egnet for bruk på de minste systemene, hvor krav til beregningseffektivitet er høye. Vi har derfor undersøkt hvorvidt denne metoden ogs˚a kan være egnet i sanntids brukerstøttesystemer for håndholdte ultralydskannere. Eksempler på anvendelser av håndholdt ultralyd innen ekkokardiografi, vil være˚a vurdere venstre hovedkammers (venstre ventrikkels) pumpefunksjon og å oppdage økning i hjertemuskemasse (hypertrofi). Venstre ventrikkels pumpefunksjon har vi vurdert ved hjelp av en algoritme som m˚aler hvor mye klaffeplanet i hjertet beveger seg opp og ned gjennom en hjertesyklus. Sykelig hypertrofi innebærer at hjertets muskelmasse vokser, uten at det pumper bedre. Vi har utviklet og testet en algoritme for ˚a oppdage hypertrofi ved å måle tykkelsen av ventrikkel septum, som er skilleveggen mellom hjertets høyre og venstre hovedkammer. Begge algoritmene synes ˚a ha en nøyaktighet god nok til en hurtig undersøkelse, men de erstatter ikke en erfaren kardiolog med en ordinær skanner. For å slippe å bruke elektrokardiogram-elektrodene som finnes på større skannere, har vi utviklet en algoritme som automatisk finner ut hvor lang hjertesykelen er, og i tillegg, når den starter. Alle algoritmene over betinger at man har et godt opptak av hjertet for analysen. Vi har anvendt Kalmanfilteret til å lage et system for å gjenkjenne standardbilder av hjertet. En videreutvikling av dette, er en assistent for ˚a hjelpe brukeren til å ta et av de vanligste og mest informative standardsnittene, det apikale firekammerbildet.
Formålet med alle disse algoritmene er å gjøre det enklere, spesielt for ikke-eksperter, å ta gode ultralydbilder av hjertet og hente ut nyttig kvantitativ informasjon ved hjelp av håndholdte ultralydsystemer. Våre resultater tilsier at Kalmanfilterbasert segmentering kan være et viktig bidrag til brukerstøttesystemer for håndholdte skannere.Medical ultrasound is widely used within the field of cardiology. The pocket-sized ultrasound scanners with performance suitable for echocardiography represent one of the latest developments within ultrasound technology. It is forecasted that these scanners will be operated also by users with less experience than what has been the case for full-size scanners. Additionally, pocket-sized devices have less possibilities for user interaction due to the reduced screen size and limited space for buttons and sliders. It is likely that these scanners would benefit from having algorithms aiding the user to capture and interpret the images.
For 3D cardiac segmentation, an efficient Kalman filtering based approach has previously been developed and shown to operate in real-time on a commercial scanner. By reducing the problem to two dimensions, this approach could probably be used to provide real-time segmentation also on pocket-sized scanners. We have developed and tested Kalman filter based solutions for improved image acquisition and quantification of some selected parameters related to left ventricle size and performance. The algorithms are intended for use with pocket-sized ultrasound devices.
The main contributions of this thesis are:
• An automatic method for measuring the mitral annulus excursion, using a combination of low frame rate speckle tracking and model based segmentation.
• A novel method for semiautomatic measurement of interventricular septal thickness using a model based Kalman filter approach.
• A timing algorithm for estimation of cardiac cycle length and cycle start without using electrocardiography (ECG).
• Methods for recognizing standard apical echocardiographic views and assisting the user to find the apical four-chamber view during acquisition.
The mitral annulus excursion and septal thickness measurement projects were evaluated using patient ultrasound recordings and manual cardiologist measurements as reference values. In case of mitral annulus excursion measurements, the accuracy of the automatic algorithm was comparable to results published by other researchers. Regarding measurements of septal thickness, the algorithm accuracy was comparable to the discrepancy between two cardiologists. Although designed for automatic operation, the algorithm turned out to frequently require manual interaction due to unsatisfactory automatic initialization. The timing algorithm robustly identified cardiac cycle length in more than 91% and cycle start in more than 77% of the 11866 test cases extracted from patient data. The view detection method was tested on 37 patient recordings and correctly classified 87% of the standard views. Using the scan assistant, ten medical students managed to capture clinically acceptable apical four-chamber recordings in 85% of the cases, compared to 55% without the scan assistant.PhD i medisinsk teknologiPhD in Medical Technolog
Analysis of the cross-section of equity returns on the JSE Securities Exchange based on linear and nonlinear modeling techniques
No full textIncludes bibliographical references.This research investigates the relationship between firm-specific style attributes and the cross-section of equity returns on the JSE Securities Exchange (JSE) over the period from 1 January 1997 to 31 December 2007. Both linear and nonlinear expected returns forecasting models are constructed based on the cross-section of equity returns. A blended approach combining a linear modeling technique with a nonlinear artificial neural network technique is developed to identify future potential top performing shares on the JSE. 1. Both linear and nonlinear models identify book-value-to-price and cash flow-to-price as significant styles attributes that distinguish near-term future share returns on the JSE. 2. This thesis found updating the identity of attributes is equally important as updating the factor payoffs of attributes in applying the stepwise regression approach. 3. Nonlinearity on the JSE equity returns is found to complement the forecasting power of linear factor models. 4. In terms of artificial neural network modeling, the extended Kalman filter learning rule introduced in the thesis is found to outperform the traditional back-propagation approach. 5. This thesis found that updating the identity of attributes via a genetic algorithm in the nonlinear forecasting models is superior to the static nonlinear forecasting models. 6. Both linear and nonlinear models are found to be more adequate in identifying future outperformers than identifying future underperformers on the JSE. The results of the research provide for potential alpha generating stock selection techniques for active portfolio managers in the South African equity market using the blended linear-nonlinear approach
