Pattern theory, a mathematical framework for representing knowledge of complex patterns developed by applied mathematician Ulf Grenander, has been shown to have potential uses in automatic target recognition (ATR). Prior research performed in the mid-1990s at Washington University in St. Louis resulted in ATR algorithms based on concepts in pattern theory for forward-looking infrared (FLIR) and laser radar (LADAR) imagery, but additional work was needed to create algorithms that could be implemented in real ATR systems. This was due to performance barriers and a lack of calibration between target models and real data. This work addresses some of these issues by exploring techniques that can be used to create practical pattern-theoretic ATR algorithms. This dissertation starts by reviewing the previous pattern-theoretic ATR research described above and discussing new results involving the unification of two previously separate outcomes of that research: multi-target detection/recognition and thermal state estimation in FLIR imagery. To improve the overall utility of pattern-theoretic ATR, the following areas are re-examined: 1) generalized diffusion processes to update target pose estimates and 2) the calibration of thermal models with FLIR target data. The final section of this dissertation analyzes the fundamental accuracy limits of target pose estimation under different sensor conditions, independent of the target detection/recognition algorithm employed. The Cramér-Rao lower bound (CRLB) is used to determine these accuracy limits.

This study’s aim is to reveal statistically significant hot spots and temporal patterns of property crime in the City and County of San Francisco and to also analyse the relationship between property crimes and the environment in which these crimes occur by using Geographic Information System (GIS). Crime pattern theory was used as the framework for the analysis of environmental surroundings and occurrence of crime. This theory indicates that certain places can be crime generators and attractors. The result showed that there are hot spots of crime in the north-eastern part of San Francisco, and that crime in these high-risk areas are intensifying. Then, by visual examination of density maps of property crime and facilities, such as shopping centres, pubs/bars/nightclubs and Bay Area Rapid Transit stations, it is shown that these facilities can explain concentrations of crime in certain areas. Furthermore, this study shows GIS can be a practical tool to utilize when presenting data of crime when used in combination with social theories which focuses on the causes of crime occurrence.

An examination of throwing was performed in a controlled environment with the aim of identifying the control and order parameters of throwing as proposed by dynamic pattern theory. A pilot study was conducted to test the possibility that the mass of a ball, the distance thrown and the size of targets were the control parameters. Based upon the results of the pilot study, only the distance was manipulated as an independent variable in the principal study.Three-dimensional motion was recorded using three video cameras in the motion analysis laboratory and later analysed using the Peak motion analysis system (software version 5.0, 1992). Sixteen right handed adult females, aged 18 - 35 years, volunteered to participate in the principal study. Subjects were seated with their trunks secured to the back support of an adjustable chair. Ten different targets (0.6 to 6.91 m) were labelled on the floor in front of the subjects. A large area for each target was defined so that the throwing skill of subjects could be ignored as a factor in the research design. Subjects were asked to throw a 0.5 kg ball to ten different distances using their own styles which allowed them to change the pattern of throwing as the distance thrown increased.All 16 subjects selected either an overarm or an underarm throw or employed both patterns. No subject used other patterns of throwing. At the shortest distance, a greater number of subjects selected an underarm throw. As the distance thrown increased, some subjects switched to the overarm throw. At the distance of 3.36 m, there were eight subjects (50%) using each style of throwing. Alteration of the throwing pattern mainly occurred from the underarm to the overarm throw. The results suggest that the distance thrown may be one of the control parameters in the throwing movement.Furthermore, the presence of both throwing patterns for all distances ++ / thrown suggests the presence of a multiple stable state in throwing motions. Trajectories of movement become more uniform as the distance thrown increased. Variability was greatest when subjects threw to the shortest distance for both patterns. These results imply that as the distance thrown increased the stability of both throwing patterns increased. Moreover, these results also imply a phase transition within each throwing pattern, in addition to the phase shift between the pattern of throwing.No result could directly illustrate the period of the transition. This may be due to the fact that phase transition in a multistable system is the result of an external force which drives the system from one state to another. Alteration of the pattern does not occur as a result of loss of the stability of the previous state. Furthermore, the presence of both throwing patterns for all distances thrown suggests the presence of a multiple stable state in throwing motions. Trajectories of movement become more uniform as the distance thrown increased. Variability was greatest when subjects threw to the shortest distance for both patterns. These results imply that as the distance thrown increased the stability of both throwing patterns increased. Moreover, these results also imply a phase transition within each throwing pattern, in addition to the phase shift between the pattern of throwing. No result could directly illustrate the period of the transition. This may be due to the fact that phase transition in a multistable system is the result of an external force which drives the system from one state to another. Alteration of the pattern does not occur as a result of loss of the stability of the previous state. However, some of the results such as the hysteresis graph presented indirect evidence, which could imply a phase shift between throwing patterns. In addition the higher ++ / degrees of joint angle recruitment in the overarm throw suggest that the stability of the system may be better maintained in the overarm throw than in the underarm throw.Alteration of the sub-styles of throwing within the same throwing pattern seemed to occur in between the shortest and the longest distances thrown. Many of the results supported this concept, for example, data related to the relative timing, the total ROM, the releasing joint angles, the trajectories of the movement, the phase plane plots, the angle-angle plots, and relative phase. However, the presence of sub-styles in the underarm throw was not as obvious as was the case for the overarm throw.In conclusion, the changes in motor behaviour during throwing as the distance thrown increased as examined in the present study can be explained by dynamic pattern theory in some aspects. However, further investigation of the stability of the patterns and energy utilisation necessary for the execution of the underarm and overarm throw are essential to determine the most suitable order parameter and to confirm the proposed control parameter (distance thrown) identified.

Mathematical and statistical modeling and analysis of biological growth using images collected over time are important for understanding of normal and abnormal development. In computational anatomy, changes in the shape of a growing anatomical structure have been modeled by means of diffeomorphic transformations in the background coordinate space. Various image and landmark matching algorithms have been developed for inference of large transformations that perform image registration consistent with the material properties of brain anatomy under study. However, from a biological perspective, it is not material constants that regulate growth, it is the genetic control system. A pattern theoretic model called the Growth as Random Iterated Diffeomorphisims (GRID) introduced by Ulf Grenander (Brown University) constructs growth-induced transformations according to fundamental biological principles of growth. They are governed by an underlying genetic control that is expressed in terms of probability laws governing the spatial-temporal patterns of elementary cell decisions (e.g., cell division/death). This thesis addresses computational and stochastic aspects of the GRID model and develops its application to image analysis of growth. The first part of the thesis introduces the original GRID view of growth-induced deformation on a fine time scale as a composition of several, elementary, local deformations each resulting from a random cell decision, a highly localized event in space-time called a seed. A formalization of the proposed model using theory of stochastic processes is presented, namely, an approximation of the GRID model by the diffusion process and the Fokker-Planck equation describing the evolution of the probability density of seed trajectories in space-time. Its time-dependent and stationary numerical solutions reveal bimodal distribution of a random seed trajectory in space-time. The second part of the thesis considers the growth pattern on a coarse time scale which underlies visible shape changes seen in images. It is shown that such a "macroscopic" growth pattern is a solution to a deterministic integro-differential equation in the form of a diffeomorphic flow dependent on the GRID growth variables such as the probability density of cell decisions and the rate of contraction/expansion. Since the GRID variables are unobserved, they have to be estimated from image data. Using the GRID macroscopic growth equation such an estimation problem is formulated as an optimal control problem. The estimated GRID variables are optimal controls that force the image of an initial organism to be continuously transformed into the image of a grown organism. The GRID-based inference method is implemented for inference of growth properties of the Drosophila wing disc directly from confocal micrographs of Wingless gene expression patterns.

Mathematical and statistical modeling and analysis of biological growth using images collected over time are important for understanding of normal and abnormal development. In computational anatomy, changes in the shape of a growing anatomical structure have been modeled by means of diffeomorphic transformations in the background coordinate space. Various image and landmark matching algorithms have been developed for inference of large transformations that perform image registration consistent with the material properties of brain anatomy under study. However, from a biological perspective, it is not material constants that regulate growth, it is the genetic control system. A pattern theoretic model called the Growth as Random Iterated Diffeomorphisims (GRID) introduced by Ulf Grenander (Brown University) constructs growth-induced transformations according to fundamental biological principles of growth. They are governed by an underlying genetic control that is expressed in terms of probability laws governing the spatial-temporal patterns of elementary cell decisions (e.g., cell division/death). This thesis addresses computational and stochastic aspects of the GRID model and develops its application to image analysis of growth. The first part of the thesis introduces the original GRID view of growth-induced deformation on a fine time scale as a composition of several, elementary, local deformations each resulting from a random cell decision, a highly localized event in space-time called a seed. A formalization of the proposed model using theory of stochastic processes is presented, namely, an approximation of the GRID model by the diffusion process and the Fokker-Planck equation describing the evolution of the probability density of seed trajectories in space-time. Its time-dependent and stationary numerical solutions reveal bimodal distribution of a random seed trajectory in space-time. The second part of the thesis considers the growth pattern on a coarse time scale which underlies visible shape changes seen in images. It is shown that such a "macroscopic" growth pattern is a solution to a deterministic integro-differential equation in the form of a diffeomorphic flow dependent on the GRID growth variables such as the probability density of cell decisions and the rate of contraction/expansion. Since the GRID variables are unobserved, they have to be estimated from image data. Using the GRID macroscopic growth equation such an estimation problem is formulated as an optimal control problem. The estimated GRID variables are optimal controls that force the image of an initial organism to be continuously transformed into the image of a grown organism. The GRID-based inference method is implemented for inference of growth properties of the Drosophila wing disc directly from confocal micrographs of Wingless gene expression patterns.

Geografisk profilering har blivit en av de mest kontroversiella och modernametoderna som används under brottsutredningar i nuläget. Framgången ochtillförlitligheten av metoden är ett debatterat ämne inom forskningsvärlden. Denhär studien ämnar att undersöka huruvida geografisk profilering är ett användbartverktyg för Polisen. Syftet med studien är att analysera hur väl metoden fungerarsom ett verktyg och komplement till en brottsutredning samt om geografiskprofilering är användbart inom bostadsinbrottutredningar. Genom att skapa ensystematisk litteraturöversikt och utföra nyckelpersonsintervjuer fann författarenatt geografisk profilering fungerar som ett utmärkt komplement till utredningar.Resultatet visade att de geografiska profileringsprogrammen inte alltid är merframgångsrika än andra metoder inom området men att de oftast är konsistenta itillförlitligheten. Resultatet visade även att metoden är användbar inombostadsinbrottutredningar så länge profilen är gjord ordentligt och utav enutbildad analytiker. Studiens slutsats är att geografisk profilering är mycket merän bara en gissningslek och kan identifiera gärningsmän om analysen är gjord aven erfaren utredare. Detta resultat diskuteras senare i studien, samt valet av metodoch möjligheter för framtida forskning. / Geographic profiling is considered as one of the most controversial andinnovative technologies used in criminal investigations today. The accuracy of themethodology has become a popular topic amongst scholars and has caused aheated debate regarding the success of geographic profiling. This study seeks toevaluate if geographic profiling is a useful tool for the police. Thus the aims ofthis study are to examine if the methodology is a viable tool during investigationsand further to establish to what extent geographic profiling has been successfullyapplied within the area of property crime, in particular burglary investigations. Byconducting a systematic literature review and key informant interviews this studyfound that geographic profiling can be a very useful tool for analysts. Further theresults showed that geographic profiling systems are not always more accuratethan simpler methods, however simpler strategies are not necessarily as consistentas a computerised system. Moreover the results indicate that geographic profilingcan be applied during burglary investigations, if done correctly and by a trainedinvestigator. The study concludes that geographic profiling is more than just aguessing game and if applied appropriately it will most likely identify theoffender. Lastly the results and shortcomings of this study, including the need forfuture research is discussed.

Studies about hot spots of crimes have found that crimes are clustered; few places have many crimes. There is a consensus among criminologists that opportunities for crimes are important when explaining hot spots, at some places, there are more opportunities than at other places. The same applies for hot spots of robberies. Most studies done on the subject are quantitative, relatively little is done using a qualitative approach. Furthermore, little research is done in a Swedish or Scandinavian context. To fill these research gaps this study use participant observations to research five hot spots of robberies in Malmö. The research will try to answer which characteristics are important to explain why the places are hot spots and what the similarities and differences there between the places are. This will be analyzed using the Routine Activity Theory and the Crime Pattern Theory. The findings suggest that place-specific things are important to explain why the places are hot spots, but when using the theories several places are similar.

According to current scholarship on offender decision making, choosing to rob another is based on a variety of individual and situational characteristics. Explanatory models often invoked within environmental criminology include routine activity, rational choice and crime pattern theories. Situational action theory’s suggestion that this decision depends, at least in part, on the interaction between offender criminal propensity and the setting’s moral context has yet to be examined. This investigation tests this idea by conducting structured interviews with active probationers and parolees centered on their decoding of streetscapes to clarify offenders’ perceptions of street robbery opportunities (Part I). These results inform an agent-based simulation contrasting the merits of assumptions made in the previously stated theories to learn how well each generates realistic concentrations of street robbery (Part II). Support emerges for both environmental criminology and situational action theory, but the results differed by the method employed. Implications follow for clarifying the theoretical processes driving these incidents and for promoting public safety. / Criminal Justice

Description of human activities in videos results not only in detection of actions and objects but also in identification of their active semantic relationships in the scene. Towards this broader goal, we present a combinatorial approach that assumes availability of algorithms for detecting and labeling objects and actions, albeit with some errors. Given these uncertain labels and detected objects, we link them into interpretative structures using domain knowledge encoded with concepts of Grenander’s general pattern theory. Here a semantic video description is built using basic units, termed generators, that represent labels of objects or actions. These generators have multiple out-bonds, each associated with either a type of domain semantics, spatial constraints, temporal constraints or image/video evidence. Generators combine between each other, according to a set of pre-defined combination rules that capture domain semantics, to form larger structures known as configurations, which here will be used to represent video descriptions. Such connected structures of generators are called configurations. This framework offers a powerful representational scheme for its flexibility in spanning a space of interpretative structures (configurations) of varying sizes and structural complexity. We impose a probability distribution on the configuration space, with inferences generated using a Markov Chain Monte Carlo-based simulated annealing algorithm. The primary advantage of the approach is that it handles known computer vision challenges – appearance variability, errors in object label annotation, object clutter, simultaneous events, temporal dependency encoding, etc. – without the need for a exponentially- large (labeled) training data set.

Till följd av att cannabisförsäljningen tillfört konflikter och varit en grundläggande orsak till ökat antal dödsskjutningar i Sverige de senaste åren, råder det därmed ingen tvekan om att det krävs satsningar för att minska cannabisförsäljningen och dess skadliga konsekvenser. Dessutom handlar en del av de globala hållbarhetsmålen i FN:s Agenda 2030 om att främja den sociala hållbarheten genom att bland annat öka tryggheten och säkerheten i den byggda stadsmiljön. Denna studie syftar på att undersöka de platser som attraherar cannabisöverlåtelse, dvs. langning, i stadsmiljön, för att kartlägga när, var och vid vilken situation som brottet förekommer. Detta för att sedan kunna ge förslag till förbättringar för utformandet av platsen för att förhindra cannabisöverlåtelse och skapa en trygg miljö för alla. I och med detta syftar arbetet till att få en ökad förståelse kring sambandet mellan säkerhet och den fysiskt utformade stadsmiljön. För att kunna uppnå syftet med studien grundar sig den kvalitativa formen av studien på data från svenska polismyndigheten där ett urval på 25 brottsplatser i Stockholms län undersöktes under kvällstid. Väl på plats fylldes en mall i löpande, kallad Trygga Platser Mallen, samtidigt som områdena fotograferades. Vidare sammanställdes och analyserades resultatet från fallstudien med underlag av olika trygghets- och säkerhets principer samt teorier. Denna studie är avgränsad och berör endast 25 brottsplatser inom Stockholmsregionen där samtliga cannabisöverlåtelser skett under kvälls-och nattetid mellan klockan 18–01.  Resultatet av studien visar att det är många miljöfaktorer som samspelar på en miljö där cannabislangning sker. Den visar att de omständigheterna på kvällstid som möjliggör cannabisöverlåtelse är i huvudsak att det blir svårare att synas, vilket gör det lättare att sälja cannabis på en del platser. Dessutom stänger de flesta verksamheter under kvällen, vilket minskar flödet kraftigt och den naturliga övervakningen försämras. På mikronivå, visar resultatet att de fysiska egenskaper som underlättar utförandet av cannabislangning är vägar, gångvägar och grönska. Vägarna kan öka tillgängligheten och mobiliteten till platsen. Dessutom ger gångvägar även en möjlighet för möte mellan potentiella kunder och cannabisförsäljare. Vidare blir grönskan ett hinder för den formella och naturliga övervakningen. Dock finns det förbättringar som kan tillämpas på den fysiska miljön för att förhindra cannabislangning. Men i huvudsak skulle det krävas en kombination mellan brottsprevention genom urban design och specifika lösningar som riktas mot en förbättring av den sociala sammanhållningen i olika områden, vilket kan vara att fler offentliga mötes-och aktivitetsplatser skapas under dagens alla timmar. I grund och botten krävs det djupare kunskap av dessa platser som är särskilt utsatta för droghandling för att kunna skapa säkrare miljöer för samtliga medborgare i dagsläget och framöver. Kandidatarbetet avslutas med en kritisk diskussion av metoden och rekommendationer för framtida studier presenteras. / In the past few years, dealing cannabis has led to conflicts, and had a significant role in the increased numbers of fatal shootings in Sweden. Hence, measures to solve the issue are undoubtedly needed to prevent these consequences. Additionally, part of the global sustainability goals in the UN's Agenda 2030 is about promoting social sustainability by, among other things, increasing the safety and security of the urban environment. The aim of this study is to examine the places in the urban environment that attract the dealing of cannabis to map when, where, and in what situation the crime occurs. Consequently, the possibility of identifying different changes to the environmental design enables the prevention of crime and promotion of safety.  To achieve the purpose of the study, a sample of 25 crime scenes in Stockholm County were examined during the evening and night, based on data from the Swedish police authority. Once in place, a template, called “Trygga Platser Mallen” (Safe Places Template), was filled in on an ongoing basis, while the physical environment was photographed at the same time. The results were then compiled and analyzed based on various safety and security principles, as well as theories. The study is delimited and therefore only contains 25 crime scenes in the Stockholm region. Also, the cannabis exchange at those places occurred during the evening between 6 pm and 1 am.  The results of the study show that there are many environmental factors that interact in a setting where dealing cannabis takes place. In essence, cannabis crime is strongly linked to the social interaction between people. However, the physical aspect facilitates the execution of the crime. The results show that the conditions of the environment in the evenings enable cannabis dealing because dealers are less likely to be seen, making it easier to sell. Additionally, most businesses close during the evening, which greatly reduces the flow of people, and the natural surveillance deteriorates. At a micro level, the results show that the physical properties that facilitate cannabis dealing are streets, walking paths, and greenery. The reason is most likely due to the streets increasing the accessibility and flow to these places. Moreover, walking paths provide an opportunity for cross paths between potential customers and dealers. Furthermore, the greenery obstructs formal and informal surveillance.  However, there are approaches to prevent the dealing of cannabis, such as combining methods for crime prevention through urban design and focusing on the improvement of social cohesion in different areas. An example would be to create spaces that encourage social interactions and activities at all hours of the day. Essentially, it is necessary to gain a deeper knowledge of the places that attract the dealing in cannabis to create safer spaces that discourages crime. Lastly, this project ends with a critical discussion of the methods and mentions recommendations for future studies.