| 1. |
- Afghahi, Morteza, 1950-
(författare)
-
Clocking of high speed CMOS VLSI systems
- 1991
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Two consequences of high level circuit integration may be increasing cost of design and cost of interconnections. Interconnection is expensive in terms of silicon area, speed and power. As far as timing is concerned. schemes of future VLSI systems may not be a simple extension of those used in the existing LSI circuits. This is because the relative importance of clock skew increases as the MOS technology develops and interconnections become slower and slower. Clock skew makes the design of VLSI synchronous circuits ineffective, complicated and failure prone. For a synchronous scheme to be useful in the VLSI environment and enjoy the experiences gained by designers, it must develop to a fast, simple, structured and robust scheme.To alleviate the adverse effects of clock skew, technological as well as circuit techniques may be employed. We have proposed a technological solution. It is suggested that a special interconnection metal layer should be introduced into the VLSI process and used for long interconnections. It is shown that by using this technique, the interconnection delay will not be a limiting factor for the performance of synchronous systems.We have also considered circuit solutions. To this end, physical causes of clock skew are investigated. It is shown that even for optimized interconnections, traditional modes of clocking results in unacceptable time performance for high speed synchronous systems. Then a new mode of clocking is presented and analysed in detail. By using this mode of clocking, the performance of synchronous systems scales with scaling the minimum feature size of MOS transistors.We have developed a synchronous scheme that is structured, simple and general. These factors also make the CMOS systems well suited for design compilation. A circuit technique is proposed that makes the design of synchronous schemes robust. Performance of different asynchronous schemes in VLSI environment is also investigated. It is shown that synchronous schemes outperforms standard asynchronous schemes for a wide range of important applications.Finally, in order to test some of the developed rules and principles, a chip has been designed as an example. In this design, a new hardware algorithm is presented for sorting. This algorithm is based on bit-serial data processing. It is shown that this design can operate at a clock frequency determined by the computational module delays and not by the clock skew.
|
|
| 2. |
- Andersson, Mats T., 1960-
(författare)
-
Controllable Multi-dimensional Filters and Models in Low-Level Computer Vision
- 1992
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- This thesis concerns robust estimation of low-level features for use in computer vision systems. The presentation consists of two parts.The first part deals with controllable filters and models. A basis filter set is introduced which supports a computationally efficient synthesis of filters in arbitrary orientations. In contrast to many earlier methods, this approach allows the use of more complex models at an early stage of the processing. A new algorithm for robust estimation of orientation is presented. The algorithm is based on synthesized quadrature responses and supports the simultaneous representation and individual averaging of multiple events. These models are then extended to include estimation and representation of more complex image primitives such as as line ends, T-junctions, crossing lines and curvature. The proposed models are based on symmetry properties in the Fourier domain as well as in the spatial plane and the feature extraction is performed by applying the original basis filters directly on the grey-level image. The basis filters and interpolation scheme are finally generalized to allow synthesis of 3-D filters. The performance of the proposed models and algorithms is demonstrated using test images of both synthetic and real world data.The second part of the thesis concerns an image feature representation adapted for a robust analogue implementation. A possible use for this approach is in analogue VLSI or corresponding analogue hardware adapted for neural networks. The methods are based on projections of quadrature filter responses and mutual inhibition of magnitude signals.
|
|
| 3. |
- Andersson, Sören
(författare)
-
On Dimension Reduction in Sensor Array Signal Processing
- 1992
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- During the last decades, sensor array signal processing has been a very active research area. More recently, relations between many of the proposed methods has been examined. The problem of assessing the estimation accuracy of these methods has also been addressed. Realworld applications of these techniques involves spatial distribution of several sensors to be used for collecting measurements of interesting emitted waveforms. From the measurements, detection and localization as well as estimation of the emitted waveforms can be accomplished. Common examples of applications are radar (electromagnetic waveforms) and sonar (acoustical underwater waveforms).Another aspect of array processing that recently has been addressed in the literature is that of dimension reduction, where the data vectors collected at the sensor outputs are reduced in size. This reduction is employed mainly in order to lower the amount of computations necessary for obtaining the parameter-estimates of interest; hut some other improvcments has also been observed. These include, e.g., lower sensitivity to sensor noise correlations and, for some estimation methods, higher resolution capability.In this thesis, it is demonstrated how to make the dimension reduction in an optimal fashion, where the optimality is with respect to estimation accuracy. More precisely, an expression to be satisfied by a transformation matrix acting on the sensor outputs is derived , that preserves the optimally achievable estimation accuracy (the Cramer-Rao bound) also in the reduced space. A transformation matrix design method that tries to reduce some unwanted properties of the optimal transformation is also outlined and examined. This method is based on numerical optimization of a particular performance mea.sure, motivated by the insight obtained in the process of finding the optimal transformation.l\foreover, an asymptotic analysis is performed, using the reduced data vectors, that examines the estimation accuracy of several estimation methods when a !arge number of sensor elements is used. This analysis is valid for a fairly general transformation matrix, and the methods considered are the Weighted Subspace Fitting (WSF) and Noise Subspace Fitting (NSF) methods, including MUSIC. By employing the optimal transformation matrix, the WSF method is shown to to be efficient, i.e., to attain the Cramer-Rao bound. An examination of the estimation accuracy, compared to that optimally attainable, is performed for the case when the transformation matrix differs from the optimal one. Finally, an application is studied, considering the potential use of sensor arrays in mobile communication systems.
|
|
| 4. |
- Babic, Ankica, 1960-
(författare)
-
Medical knowledge extraction : application of data analysis methods to support clinical decisions
- 1993
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- In building computer based clinical decision support extensive data analysis is sought to acquire all the medical knowledge needed to formulate the decision rules.This study explores, compares and discusses several approaches to knowledge extraction from medical data. Statistical methods (univariate, multivariate), probabilistic artificial intelligence approaches (inductive learning procedures, neural networks) and the rough sets were used for this purpose. The methods were applied in two clinical sets of data with well defined patients groups.The aim of the study was then to use different data analytical methods and extract knowledge, both of semantic and classification nature, enabling to differentiate among patients, observations and disease groups, what in turn was aimed to support clinical decisions.Semantic analysis was performed in two ways. In prior analysis subgroups or patterns were formed based on the distance within the data, while in posterior semantic analysis 'types' of observation falling into various groups and their measured values were explored.To study further discrimination, two empirical systems, based on principles of learning from examples, i.e. based on Quintan's ID3 algorithm (the AssPro system) and CART (Classification and Regression Trees), were compared. The knowledge representation in both systems is tree structured, thus the comparison is made according to the complexity, accuracy and structure of their optimal decision trees. The inductive learning system was additionaly compared and evaluated in relation to the location model of discriminant analysis, the linear Ficherian discrimination and the rough sets.All methods used were analysed and compared for their theoretical and applicative performances, and in some cases they were assessed medical appropriateness. By using them for the extensive knowledge extraction, it was possible to give a strong methodological basis for design of clinical decision support systems specific for the problem and the medical environments considered.
|
|
| 5. |
- Birch, Jens
(författare)
-
Single-crystal Mo/V superlattices : growth, structure, and hydrogen uptake
- 1994
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Fundamental studies concerning the growth, structural characterization and hydrogen uptake of single-crystal (00 l )-oriented Mo/V superlattices have been performed. The superlattices were grown by dual-target magnetron sputtering in pure Ar-atmosphere < 6·10-3 Torr on (001)-oriented MgO substrates. X-ray diffraction (XRD), X-ray and neutron reflectivity, high resolution (HR) as well as ordinary crosssectional transmission electron microscopy (XTEM) and selected area electron diffraction (SAED) were used for the structural characterization. Hydrogen depth-profiling was performed by the 15N method.For growth of periodic Mo/V superlattices, it is shown that substrate temperatures in the range of 600-700 °C is feasible for epitaxy. At higher growth temperatures substantial interdiffusion occurred. Furthermore, simulations of XRDpatterns gave the width of the interfaces to be ±1 monolayer (±0,154 nm) which was confirmed by XRD and HRXTEM analyses of a superlattice grown with layer thicknesses DMo=Dv=0,31 nm (2 monolayers). A transition from smooth to wavy V-layers was found to occur at a critical V-layer thickness Dc. In superlattices where the relative amount of V is large, De is large and vice versa for superlattices containing thin V-layers. In superlattices with equally thick Mo- and V-layers Dc was found to be ~2,5 nm. Mo was found to grow with a uniform thickness following the surface of the V-layers. The layer thickness fluctuations are non-accumulative and disappear if the periodicity of a growing Mo/V superlattice is changed so that Dv becomes smaller than Dc. The origin of the 3D evolution is explained in terms of surface strain and the roughening transition. The interfaces of Mo/V superlattices grown under the influence of energetic ion bombardment ranging from about 15 eV to 250 eV was studied by HRXTEM and XRD. Both techniques indicated a continous deterioration of the interface quality and an increasing amount of defects with increasing ion energy.The diffraction peaks from a clas of quasi-periodic superlattices which can be generated by the inflation rules A→AmB, B→A (m = positive integer) was analytically, experimentally and numerically found to be located at the wavevectors q = 2πɅ-1rγ(m)k where r and k are integers and A is an average superlattice period. The ratios, γ(m), between the thicknesses of the two superlattice building blocks, A and B, must be chosen such that γ(m) = (m + (m2 + 4) 1/2 )/2.The uptake of hydrogen in the superlattices is found to decrease with decreasing A and for ≤5,5 nm the transition between α-VHx and β-VHx is not observed. A model is proposed which explains the A-dependent behaviour of the hydrogen uptake by a transfer of interstitial electrons from Mo to V, creating a 0,49 nm wide H-free interface layer. The existence of this layer is shown both by the 15N method performed on samples containing several A:s and by combining simulations of X-ray and neutron reflectivities with measurements on superlattices loaded with either hydrogen or deuterium. The structural change of Mo/V(OOl) superlattices upon H-loading was measured by a method derived in this work which utilises a combination of X-ray reflectivity and reciprocal space mapping by XRD. The lattice parameters in the layers are measured in the growth direction as well a in the plane of the sample. It is found that the V lattice expands in the growth direction and that the hydrogenation process is associated with relaxation of coherency strain.
|
|
| 6. |
- Björn, Anders, 1966-
(författare)
-
Removable singularities for Hardy spaces of analytic functions
- 1994
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Let Ω ⊂ C be a non-empty open connected set and O < p < ∞. Define Hp(Ω) to be the set of all analytic functions f in Ω such that there exists a harmonic function uf in Ω with |f|p < uf· Let K ⊂ Ω be compact and such that Ω ∖, K is connected. Then the set K is said to be a removable singularity for Hp(Ω ∖ K) if Hp(Ω)= Hp(Ω ∖ K). Hejhal proved in 1973 that this notion does not depend on Ω.In this thesis we give a survey of the theory of removable singularities for Hardy spaces. We use potential theory, conformal mappings, harmonic measures and Banach space techniques to give new results. One new result is that if dim K > min {1, p} then K is not removable. Several theorems about removability of sets lying on rectifiable curves and also conditions for removability of some planar self-similar Cantor sets are given.
|
|
| 7. |
- Bårman, Håkan
(författare)
-
Hierarchical curvature estimation in computer vision
- 1991
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- This thesis concerns the estimation and description of curvature for computer vision applications. Different types of multi-dimensional data are considered: images (2D); volumes (3D); time sequences of images (3D); and time sequences of volumes (4D).The methods are based on local Fourier domain models and use local operations such as filtering. A hierarchical approach is used. Firstly, the local orientation is estimated and represented with a vector field equivalent description. Secondly, the local curvature is estimated from the orientation description. The curvature algorithms are closely related to the orientation estimation algorithms and the methods as a whole give a unified approach to the estimation and description of orientation and curvature. In addition, the methodology avoids thresholding and premature decision making.Results on both synthetic and real world data are presented to illustrate the algorithms performance with respect to accuracy and noise insensitivity. Examples illustrating the use of the curvature estimates for tasks such as image enhancement are also included.
|
|
| 8. |
- Chowdhury, Shamsul I., 1949-
(författare)
-
Computer-based support for knowledge extraction from clinical databases
- 1990
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- This thesis is devoted to aspects related to the analysis and interpretation of medical data thus allowing knowledge extraction from medical databases. The difficulties with traditional approaches for data analysis and interpretation by people less proficient in statistics are well known. An attempt has been made to identify different stages in the processes of data analysis and interpretation where this category of users might need help and to identify how this help could best be provided (studies I, II, III). In this work artificial intelligence approaches have been used as remedies to improve user-friendliness as well as power and effectiveness of statistical software in respect to data analysis strategies. Prototype implementations based on these approaches are presented and discussed. Issues pertaining to the evaluation of decision support systems m medicine have been identified and discussed in detail (study IV).Knowledge in a knowledge-based system for decision support is generally acquired from experts and the literature. Knowledge can also be effectively extracted from a database of patient observations and from interpretation of those observations. The resulting system would be more accurate in the latter case, especially if it is intended to operate in decision support in the same clinical setting.The studies V and VI were conducted to show how retrospectively collected data could be utilized for the purpose of knowledge extraction. More traditional data analysis approaches (study V) were used to analyze a database on liver diseases. The data material used in the study was collected in the HELP system as a routine part of patient care. The main issues involved were detection of outliers and treatment of missing values in order to facilitate utilization of this kind of database for eventual knowledge extraction. In Study V, statistical techniques including discriminant analysis and artificial intelligence approaches such asinductive learning, were used. The 'K nearest neighbor' technique was found to be an easy and acceptable method for estimating missing values when the database contained only a few missing values for each object in the database. Discriminant analysis was found to be a good method for classifying a patient, based on a set of variables, into two or more disease classes. The results show that when discriminant analysis was applied to two groups based on a relatively large number (19) of variables, then only a few (3) of the variables accounted for a high percentage of correct classifications.The knowledge-based approach for data analysis and interpretation used in (study III) was applied to a large database (study VI). The main emphasis was to study the feasibility of the approach in exploring a large patient record system. The data material was taken from Kronan Health Center - a primary health care center in suburban Stockholm with a patient database consisting of about 14,000 medical records. The analysis was carried out to test the hypothesis of a possible causation between hypertension and diabetes. The results of this study support the assumption that there is a relationship between diabetes and hypertension but the question of the direction of this relationship remained unsolved, as did the question of direct causality. On the other hand, the results of this study are in accordance with the hypothesis of a common metabolic syndrome. The results arrived at by the analysis method (multivariate tabular analysis) utilized by the system are, moreover in accordance with another statistical method (log linear analysis). This also supports the approach taken in the knowledge-based system.
|
|
| 9. |
- Dahlbäck, Nils, 1949-
(författare)
-
Representations of discourse : cognitive and computational aspects
- 1991
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- This work is concerned with empirical studies of cognitive and computational aspects of discourse representations. A more specific aim is to contribute to the development of natural language interfaces for interaction with computers, especially the development of representations making possible a continuous interactive dialogue between user and system.General issues concerning the relationship between human cognitive and computational aspects of discourse representations are studied through an empirical and theoretical analysis of a psychological theory of discourse coherence, the theory of mental models. The analysis suggests that there are principled limits to what workers in computational linguistics can learn from psychological work on discourse processing.As far as the theory of mental models as a psychologica1 theory of discourse is concerned, the effect of previous background knowledge of the domain of discourse on the processing of the types of texts often used in previous work is demonstrated. It is argued that this demonstration does not invalidate any of the basic assumptions of the theory, but should rather be seen as a modification or clarification. An attempt is also made to study the possible existence of different cognitive strategies used by different subjects and in different tasks.While some supporting evidence for this can be seen, it is argued that the results-obtained are not conclusive on this issue.Another set of studies use the so-called Wizard-of-Oz method, i.e. dialogues with simulated natural language interfaces. Here the focus of the analysis is on the dialogue structure, and on the use of referring and co-referring expressions in the dialogues. The basic result of thedialogue analysis is that it is possible to describe these kinds of dialogues using a dialogue grammar, the LINDA-model, the basic feature of which is the partitioning of dialogues in a number of initiative-response (IR) units. The study of referring expressions also shows a lack of some of the complexities encountered in human dialogues. The results point to the possibility of using computationally simpler methods than what has hitherto been assumed, both for the dialogue management and for the resolution of anaphoric references.
|
|
| 10. |
- Doherty, Patrick, 1957-
(författare)
-
NML3 : a non-monotonic formalism with explicit defaults
- 1991
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- The thesis is a study of a particular approach to defeasible reasoning based on the notion of an information state consisting of a set of partial interpretations constrained by an information ordering. The formalism proposed, called NML3, is a non-monotonic logic with explicit defaults and is characterized by the following features: (1) The use of the strong Kleene three-valued logic as a basis. (2) The addition of an explicit default operator which enables distinguishing tentative conclusions from ordinary conclusions in the object language. (3) The use of the technique of preferential entailment to generate non-monotonic behavior. The central feature of the formalism, the use of an explicit default operator with a model theoretic semantics based on the notion of a partial interpretation, distinguishes NML3 from the existing formalisms. By capitalizing on the distinction between tentative and ordinary conclusions, NML3 provides increased expressibility in comparison to many of the standard non-monotonic formalisms and greater flexibility in the representation of subtle aspects of default reasoning.In addition to NML3, a novel extension of the tableau-based proof technique is presented where a signed formula is tagged with a set of truth values rather than a single truth value. This is useful if the tableau-based proof technique is to be generalized to apply to the class of multi-valued logics. A refutation proof procedure may then be used to check logical consequence for the base logic used in NML3 and to provide a decision procedure for the propositional case of NML3.A survey of a number of non-standard logics used in knowledge representation is also provided. Various formalisms are analyzed in terms of persistence properties of formulas and their use of information structures.
|
|
