| 1. |
- Fredenberg, Erik, 1979-
(författare)
-
Spectral Mammography with X-Ray Optics and a Photon-Counting Detector
- 2009
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Early detection is vital to successfully treating breast cancer, and mammography screening is the most efficient and wide-spread method to reach this goal. Imaging low-contrast targets, while minimizing the radiation exposure to a large population is, however, a major challenge. Optimizing the image quality per unit radiation dose is therefore essential. In this thesis, two optimization schemes with respect to x-ray photon energy have been investigated: filtering the incident spectrum with refractive x-ray optics (spectral shaping), and utilizing the transmitted spectrum with energy-resolved photon-counting detectors (spectral imaging). Two types of x-ray lenses were experimentally characterized, and modeled using ray tracing, field propagation, and geometrical optics. Spectral shaping reduced dose approximately 20% compared to an absorption-filtered reference system with the same signal-to-noise ratio, scan time, and spatial resolution. In addition, a focusing pre-object collimator based on the same type of optics reduced divergence of the radiation and improved photon economy by about 50%. A photon-counting silicon detector was investigated in terms of energy resolution and its feasibility for spectral imaging. Contrast-enhanced tumor imaging with a system based on the detector was characterized and optimized with a model that took anatomical noise into account. Improvement in an ideal-observer detectability index by a factor of 2 to 8 over that obtained by conventional absorption imaging was found for different levels of anatomical noise and breast density. Increased conspicuity was confirmed by experiment. Further, the model was extended to include imaging of unenhanced lesions. Detectability of microcalcifications increased no more than a few percent, whereas the ability to detect large tumors might improve on the order of 50% despite the low attenuation difference between glandular and cancerous tissue. It is clear that inclusion of anatomical noise and imaging task in spectral optimization may yield completely different results than an analysis based solely on quantum noise.
|
|
| 2. |
- Hilletofth, Per
(författare)
-
Demand-Supply Chain Management
- 2010
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Purpose: This research aims to enhance the current understanding and knowledge of the demand-supply chain management (DSCM) concept by determining its elements, benefits, and requirements, as well as by analyzing key elements of the concept. Methodology: This research has utilized the case study strategy and the survey strategy, however, the case study strategy dominates. The case study research has involved five companies originating from Sweden and the collection of empirical data mainly from in-depth interviews with key persons representing senior and middle management. The survey research targeted the largest firms in Sweden and Finland and empirical data was collected through an online questionnaire. Findings: This research has established that the main elements of DSCM include market orientation, coordination of the demand and supply processes, viewing the demand and supply processes as being equally important, as well as value creation, differentiation, innovativeness, responsiveness, and cost-efficiency in the demand and supply processes. It has also been revealed that the main benefits of DSCM include enhanced competiveness, enhanced demand chain performance, as well as enhanced supply chain performance, while the main requirements of DSCM include organizational competences, company established principles, demand-supply chain collaboration, and information technology support. A key element of DSCM further investigated is differentiation focused supply chain design. It has been shown that these efforts can be organized into a process of five stages. In addition, it is important that this process is addressed in parallel with the new product development (NPD) process, that information is exchanged between them, and that they are directed on the basis of the same segmentation model. Another key element of DSCM further investigated is coordination between NPD and SCM. This research has identified several significant linkages between these management directions, which motivate the use of an integrative NPD process where the NPD functions are aligned with the main supply functions in the company and other sales-related functions supporting the commercialization. A final key element of DSCM further investigated is the significance of regarding the demand processes and the supply processes as being equally important. This research has revealed that logistics outsourcing can be risky, if it results in the supply processes being considered less important. Nevertheless, if senior management regards the outsourced processes as equally important as the in-house processes, the effect of logistics outsourcing on company strategies and direction in SCM could be reduced and logistics outsourcing could instead provide an opportunity to improve the design and differentiation of the supply chain. Research limitations/implications: This research has proposed, described, and further analyzed a demand-supply oriented management approach. Such a management approach stresses that the demand processes and the supply processes have to be coordinated and directed at an overlying level, in order to gain and sustain a competitive advantage in competitive and fragmented markets. This research is mainly explorative in nature, and more empirical data, from similar and other research settings, is needed to further validate the findings. Another limitation of the research is that it is essentially limited to Swedish companies (even if some Finnish companies are involved in the survey), however, many of the case companies have a large international presence and are among the top three in their industries, facts which provide some grounds for generalization. Practical implications: This research provides researchers and practitioners with insights into how to develop a demand-supply oriented business. It shows that companies should organize themselves around understanding how customer value is created and delivered, as well as how these processes and management directions can be coordinated. In order for this to occur, the demand and supply processes must be considered as being equally important and the firm needs to be managed jointly and in a coordinated manner by the demand- and supply-side of the company. It is also important that value creation is considered in both the demand and supply processes. Originality/value: Despite strong arguments from both researchers and practitioners for a demand-supply oriented management approach only a minority of companies appear to have effectively coordinated the demand and supply processes. This might be influenced by the lack of research examining how the demand and supply processes can be coordinated, what benefits can be gained by coordinating them, and what requirements are necessary to succeed. This research contributes by investigating these types of aspects further.
|
|
| 3. |
- Chatterjee, Bapi, 1982
(författare)
-
Lock-free Concurrent Search
- 2017
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- The contemporary computers typically consist of multiple computing cores with high compute power. Such computers make excellent concurrent asynchronous shared memory system. On the other hand, though many celebrated books on data structure and algorithm provide a comprehensive study of sequential search data structures, unfortunately, we do not have such a luxury if concurrency comes in the setting. The present dissertation aims to address this paucity. We describe novel lock-free algorithms for concurrent data structures that target a variety of search problems. (i) Point search (membership query, predecessor query, nearest neighbour query) for 1-dimensional data: Lock-free linked-list; lock-free internal and external binary search trees (BST). (ii) Range search for 1-dimensional data: A range search method for lock-free ordered set data structures - linked-list, skip-list and BST. (iii) Point search for multi-dimensional data: Lock-free kD-tree, specially, a generic method for nearest neighbour search. We prove that the presented algorithms are linearizable i.e. the concurrent data structure operations intuitively display their sequential behaviour to an observer of the concurrent system. The lock-freedom in the introduced algorithms guarantee overall progress in an asynchronous shared memory system. We present the amortized analysis of lock-free data structures to show their efficiency. Moreover, we provide sample implementations of the algorithms and test them over extensive micro-benchmarks. Our experiments demonstrate that the implementations are scalable and perform well when compared to related existing alternative implementations on common multi-core computers. Our focus is on propounding the generic methodologies for efficient lock-free concurrent search. In this direction, we present the notion of help-optimality, which captures the optimization of amortized step complexity of the operations. In addition to that, we explore the language-portable design of lock-free data structures that aims to simplify an implementation from programmer’s point of view. Finally, our techniques to implement lock-free linearizable range search and nearest neighbour search are independent of the underlying data structures and thus are adaptive to similar data structures.
|
|
| 4. |
- Ali, Muhaddisa Barat, 1986
(författare)
-
Deep Learning Methods for Classification of Gliomas and Their Molecular Subtypes, From Central Learning to Federated Learning
- 2023
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- The most common type of brain cancer in adults are gliomas. Under the updated 2016 World Health Organization (WHO) tumor classification in central nervous system (CNS), identification of molecular subtypes of gliomas is important. For low grade gliomas (LGGs), prediction of molecular subtypes by observing magnetic resonance imaging (MRI) scans might be difficult without taking biopsy. With the development of machine learning (ML) methods such as deep learning (DL), molecular based classification methods have shown promising results from MRI scans that may assist clinicians for prognosis and deciding on a treatment strategy. However, DL requires large amount of training datasets with tumor class labels and tumor boundary annotations. Manual annotation of tumor boundary is a time consuming and expensive process. The thesis is based on the work developed in five papers on gliomas and their molecular subtypes. We propose novel methods that provide improved performance. The proposed methods consist of a multi-stream convolutional autoencoder (CAE)-based classifier, a deep convolutional generative adversarial network (DCGAN) to enlarge the training dataset, a CycleGAN to handle domain shift, a novel federated learning (FL) scheme to allow local client-based training with dataset protection, and employing bounding boxes to MRIs when tumor boundary annotations are not available. Experimental results showed that DCGAN generated MRIs have enlarged the original training dataset size and have improved the classification performance on test sets. CycleGAN showed good domain adaptation on multiple source datasets and improved the classification performance. The proposed FL scheme showed a slightly degraded performance as compare to that of central learning (CL) approach while protecting dataset privacy. Using tumor bounding boxes showed to be an alternative approach to tumor boundary annotation for tumor classification and segmentation, with a trade-off between a slight decrease in performance and saving time in manual marking by clinicians. The proposed methods may benefit the future research in bringing DL tools into clinical practice for assisting tumor diagnosis and help the decision making process.
|
|
| 5. |
- Fischer, Björn
(författare)
-
A Socio-Material Study of User Involvement : Interrogating the practices of technology development for older people in a digitalised world
- 2022
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Population ageing and increased digitalization each constitute an ongoing and profound transformation within contemporary modes of living, as growing advances in technological development mix and intermingle with the lived realities of older people as the final recipients. It is against the backdrop of this interplay that user involvement has enjoyed ever-rising advocacy to an almost normative degree. Beyond articulating methodological principles, however, the literature has remained surprisingly vague as to the practical implementation of the approach. Less appears to be known, both empirically and conceptually, about how design and user involvement are done in practice and how they would matter to bring about intentional or unintentional effects. To engage with these developments, this thesis aims at taking the practices of user involvement and design to the centre of its inquiry by adopting a perspective from Science and Technology Studies (STS). Specifically, the thesis seeks to both build on and contribute to the established body of STS on the connection between technology design and older users and ask: What is there to learn about user involvement as a method, if we focus on the practices of doing user involvement? To answer this question, the thesis studies four different aspects of the practices of user involvement and design. In particular, the thesis reviews the literature on how user involvement mattered in previous empirical projects that include older people (Paper I), it examines how different configurations of participation matter in design workshops (Paper II), it scrutinizes the achievement of user involvement in corporate practices (Paper III) and it traces the circumstantial performances of such practices (Paper IV). The largest empirical piece comes from a two-year ethnographic study of a small- to medium-sized enterprise, the material from which informed Paper III and Paper IV.The findings highlight how user involvement in practice is both contingent and transformative, as it selectively enrols different participants and performs multiple realities. In practice, user involvement appears to be dependent on a set of underlying premises and socio-material conditions and thus is always a dynamic and momentary achievement. Furthermore, the thesis shows how the practices of user involvement themselves may bring into existence different realities, articulating and materializing particular versions of objects and images of ageing. Accordingly, the thesis contributes theoretically by illuminating the underlying socio-material facets of user involvement, and by emphasizing ageing as a particular object/image of design. Specifically, the appended papers encompass a conceptual framework, as well as three new concepts: design multiple, shifting interstices and viscous image landscape, in order to theorize the underlying conditions for user involvement, its relationship with design and its entanglement with ageing. Practically, the thesis enunciates three main implications regarding questions of goodness, politics and ethics.
|
|
| 6. |
- Homem, Irvin, 1985-
(författare)
-
Advancing Automation in Digital Forensic Investigations
- 2018
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Digital Forensics is used to aid traditional preventive security mechanisms when they fail to curtail sophisticated and stealthy cybercrime events. The Digital Forensic Investigation process is largely manual in nature, or at best quasi-automated, requiring a highly skilled labour force and involving a sizeable time investment. Industry standard tools are evidence-centric, automate only a few precursory tasks (E.g. Parsing and Indexing) and have limited capabilities of integration from multiple evidence sources. Furthermore, these tools are always human-driven.These challenges are exacerbated in the increasingly computerized and highly networked environment of today. Volumes of digital evidence to be collected and analyzed have increased, and so has the diversity of digital evidence sources involved in a typical case. This further handicaps digital forensics practitioners, labs and law enforcement agencies, causing delays in investigations and legal systems due to backlogs of cases. Improved efficiency of the digital investigation process is needed, in terms of increasing the speed and reducing the human effort expended. This study aims at achieving this time and effort reduction, by advancing automation within the digital forensic investigation process.Using a Design Science research approach, artifacts are designed and developed to address these practical problems. Summarily, the requirements, and architecture of a system for automating digital investigations in highly networked environments are designed. The architecture initially focuses on automation of the identification and acquisition of digital evidence, while later versions focus on full automation and self-organization of devices for all phases of the digital investigation process. Part of the remote evidence acquisition capability of this system architecture is implemented as a proof of concept. The speed and reliability of capturing digital evidence from remote mobile devices over a client-server paradigm is evaluated. A method for the uniform representation and integration of multiple diverse evidence sources for enabling automated correlation, simple reasoning and querying is developed and tested. This method is aimed at automating the analysis phase of digital investigations. Machine Learning (ML)-based triage methods are developed and tested to evaluate the feasibility and performance of using such techniques to automate the identification of priority digital evidence fragments. Models from these ML methods are evaluated in identifying network protocols within DNS tunneled network traffic. A large dataset is also created for future research in ML-based triage for identifying suspicious processes for memory forensics.From an ex ante evaluation, the designed system architecture enables individual devices to participate in the entire digital investigation process, contributing their processing power towards alleviating the burden on the human analyst. Experiments show that remote evidence acquisition of mobile devices over networks is feasible, however a single-TCP-connection paradigm scales poorly. A proof of concept experiment demonstrates the viability of the automated integration, correlation and reasoning over multiple diverse evidence sources using semantic web technologies. Experimentation also shows that ML-based triage methods can enable prioritization of certain digital evidence sources, for acquisition or analysis, with up to 95% accuracy.The artifacts developed in this study provide concrete ways to enhance automation in the digital forensic investigation process to increase the investigation speed and reduce the amount of costly human intervention needed.
|
|
| 7. |
- Löfgren, Johan, 1983
(författare)
-
Local Sea Level Observations Using Reflected GNSS Signals
- 2014
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Sea level rise due to global warming is predicted to have a large impact on human society, especially for populations living in coastal regions and on islands. It is therefore of great importance to monitor the sea level and to increase the understanding of the local hydrodynamic and meteorological responses to a global sea level rise.The focus of this thesis is to estimate the local sea level using Global Navigation Satellite System (GNSS) signals reflected off the sea surface. These signals were recorded in two different ways using a GNSS tide gauge at the Onsala Space Observatory, consisting of standard geodetic-type commercially off-the-shelf GNSS equipment. First, the phase-delay of the reflected GNSS signals were recorded directly with a receiver connected to a nadir-looking antenna. Together with the phase-delay of the direct signals, recorded with a receiver connected to a zenith-looking antenna, standard geodetic analysis provided GNSS sea level observations. Second, the Signal-to-Noise Ratio (SNR) recorded with the receiver connected to the zenith-looking antenna, provided an indirect measurement of the reflected GNSS signals, as the reflected signals interfered with the direct GNSS signals and affected the recorded observables. From analysis of the multipath oscillations, an additional type of sea level observation was possible. Furthermore, the SNR-analysis method allowed other GNSS stations, located close to the ocean, in different parts of the world to become GNSS tide gauges.The GNSS-derived sea level from the GNSS tide gauge at the observatory was compared with independent observations of sea level from co-located traditional tide gauges, showing a high level of agreement with correlation coefficients of 0.89-0.99. The sea level results from the phase-delay analysis performed better with respect to the traditional sea level records than the results from the SNR-analysis. As an example, the Root-Mean-Square (RMS) differences from 1 month of observations between the GNSS-derived sea level (using frequency band L1) and the sea level from the co-located tide gauge were 3.2-3.5 cm and 4.0-4.7 cm for the phase-delay analysis and the SNR-analysis, respectively.Sea level results applying the SNR-analysis for data of 5 different GNSS stations around the world were compared to independent co-located traditional tide gauge records. The results showed RMS differences on the order of 6.2 cm for stations with low tidal ranges (up to 165 cm) and 43 cm for stations with high tidal ranges (up to 772 cm). In this case, an extended SNR-analysis approach was applied, modelling a time dependent sea level.
|
|
| 8. |
- Svärd, Malin, 1985
(författare)
-
Computational driver behavior models for vehicle safety applications
- 2023
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- The aim of this thesis is to investigate how human driving behaviors can be formally described in mathematical models intended for online personalization of advanced driver assistance systems (ADAS) or offline virtual safety evaluations. Both longitudinal (braking) and lateral (steering) behaviors in routine driving and emergencies are addressed. Special attention is paid to driver glance behavior in critical situations and the role of peripheral vision. First, a hybrid framework based on autoregressive models with exogenous input (ARX-models) is employed to predict and classify driver control in real time. Two models are suggested, one targeting steering behavior and the other longitudinal control behavior. Although the predictive performance is unsatisfactory, both models can distinguish between different driving styles. Moreover, a basic model for drivers' brake initiation and modulation in critical longitudinal situations (specifically for rear-end conflicts) is constructed. The model is based on a conceptual framework of noisy evidence accumulation and predictive processing. Several model extensions related to gaze behavior are also proposed and successfully fitted to real-world crashes and near-crashes. The influence of gaze direction is further explored in a driving simulator study, showing glance response times to be independent of the glance's visual eccentricity, while brake response times increase for larger gaze angles, as does the rate of missed target detections. Finally, the potential of a set of metrics to quantify subjectively perceived risk in lane departure situations to explain drivers' recovery steering maneuvers was investigated. The most influential factors were the relative yaw angle and splay angle error at steering initiation. Surprisingly, it was observed that drivers often initiated the recovery steering maneuver while looking off-road. To sum up, the proposed models in this thesis facilitate the development of personalized ADASs and contribute to trustworthy virtual evaluations of current, future, and conceptual safety systems. The insights and ideas contribute to an enhanced, human-centric system development, verification, and validation process. In the long term, this will likely lead to improved vehicle safety and a reduced number of severe injuries and fatalities in traffic.
|
|
| 9. |
- Jeong, Seung Hee, 1978-
(författare)
-
Soft Intelligence : Liquids Matter in Compliant Microsystems
- 2016
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Soft matter, here, liquids and polymers, have adaptability to a surrounding geometry. They intrinsically have advantageous characteristics from a mechanical perspective, such as flowing and wetting on surrounding surfaces, giving compliant, conformal and deformable behavior. From the behavior of soft matter for heterogeneous surfaces, compliant structures can be engineered as embedded liquid microstructures or patterned liquid microsystems for emerging compliant microsystems.Recently, skin electronics and soft robotics have been initiated as potential applications that can provide soft interfaces and interactions for a human-machine interface. To meet the design parameters, developing soft material engineering aimed at tuning material properties and smart processing techniques proper to them are to be highly encouraged. As promising candidates, Ga-based liquid alloys and silicone-based elastomers have been widely applied to proof-of-concept compliant structures.In this thesis, the liquid alloy was employed as a soft and stretchable electrical and thermal conductor (resistor), interconnect and filler in an elastomer structure. Printing-based liquid alloy patterning techniques have been developed with a batch-type, parallel processing scheme. As a simple solution, tape transfer masking was combined with a liquid alloy spraying technique, which provides robust processability. Silicone elastomers could be tunable for multi-functional building blocks by liquid or liquid-like soft solid inclusions. The liquid alloy and a polymer additive were introduced to the silicone elastomer by a simple mixing process. Heterogeneous material microstructures in elastomer networks successfully changed mechanical, thermal and surface properties.To realize a compliant microsystem, these ideas have in practice been useful in designing and fabricating soft and stretchable systems. Many different designs of the microsystems have been fabricated with the developed techniques and materials, and successfully evaluated under dynamic conditions. The compliant microsystems work as basic components to build up a whole system with soft materials and a processing technology for our emerging society.
|
|
| 10. |
- Schweigler, Michael
(författare)
-
Nonlinear modeling of reinforced dowel joints in timber structures : a combined experimental-numerical study
- 2018
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Steel dowels are indispensable elements for the design of joints in modern timber structures. Dowels are broadly used because of their flexibility in design and easy assembling on-site, as well as due to their advantageous mechanical behavior. Recent developments in reinforcement techniques allow for designing ductile dowel joints, which exhibit nonlinear slip behavior. However, currently applied limit state approaches for calculation of connection strength are not able to fully exploit the potential of dowel joints. This calls for development of more advanced calculation methods, which was aimed for in this thesis.For thorough understanding of the complex mechanical behavior of dowel connections, application of a so-called multiscale approach is advantageous. Thereby, nonlinear loaddeformation behavior of dowel connections is studied on different length scales, from the scale of connection components, up to the joint level. The aim of this thesis was to exploit knowledge from lower scales in models that finally allow for nonlinear analysis of timber structures. In the work at hand, this was achieved by a combined experimental-numerical analysis.Experimental studies on the nonlinear embedment slip of ductile dowel connections, and its relation to the orthotropic material behavior of wood, was one of the objectives of this work. Breaking new ground by testing up to large dowel displacements, at various angles to the grain, for unconstrained and constrained lateral displacement conditions, required development of new testing procedures and test setups. Test results gave access to nonlinear embedment slip curves and showed their dependence on loading direction and lateral displacement conditions.In the next step, embedment slip data were exploited in modeling of single-dowel connections. Beam-on-nonlinear foundation modeling was applied for this purpose. Validation of connection tests and a parameter study not only highlighted suitability of the calculation method, but allowed for gaining insight into limit states. Thus, beam-on-foundation modeling was found to be an attractive alternative to advanced 3D FEM models for engineering design.Parameterized equations for regression analysis of nonlinear slip curves, and interaction curves describing the grain angle dependence of mechanical parameters, were summarized in a literature review. With these equations at hand, a multi-dimensional parameterization method for the nonlinear slip as a function of the load-to-grain angle was developed. This method was applied to experimental data derived in embedment and single-dowel connection tests as well as in simulations. Analytical equations for connection slip are expected to facilitate engineering modeling at the single-dowel connection and joint level, respectively.Joint modeling aimed at establishing a calculation method suitable for engineering design with an attractive trade-off between modeling effort, calculation time and accuracy. This was tackled by a semi-analytical model based on nonlinear elastic springs for the dowel slip and rigid connection members. Thereby, global joint slip, and thus stiffness and strength of joints, as well as local load distribution within joints can be predicted. This model proved to be suitable for single-dowel-based and joint-based design concepts. Calculation examples showed pronounced influence of loading direction dependence and nonlinearity of dowel slip on local load distribution, as well as on global joint slip. Influence of interaction between internal forces on the joint stiffness became obvious from structural analysis with nonlinear joint slip.This work covers a combined experimental-numerical analysis of the slip behavior of ductile dowel joints, from the wood embedment and steel dowel bending behavior, the single-dowel slip, to the joint behavior, with final application in nonlinear structural analysis. Presented models predicted nonlinear slip with suitable accuracy and efficiency. Application of nonlinear joint slip in structural analysis showed the potential of ductile dowel joints, which could be exploited in engineering design by the herein presented multiscale modeling strategy.
|
|
