· · · Institutional Repository

Object Tracking has been a very active area in the field of C omputer Vision. Over the years, a variety of approaches have been put forth to solve this problem and though many of them have demonstrate considerable success none of them have been completely successful. With more methods being written each day, the evaluation of such systems becomes a very important task. If an evaluation system exists that is able to point out specific flaws in the stage of development, it can lead to a very robust and improved algorithm. This work attempts to create such an evaluation framework. Given an algorithm that detects people and simultaneously tracks them, we evaluate its output by considering the complexity of the input scene. Some videos used for the evaluation are recorded using the Kinect sensor and a benchmark dataset from the PETS workshop is also used. To analyze the performance of the tracking system,the reasons due to which the algorithm might fail are investigated and quantified over the entire video sequence. A set of features called Scene C omplexity Measures are obtained for each input frame. The variability in the algorithm performance is modeled by these complexity measures using various regression models. From the regression statistics, we show that we can compare the performance of two different algorithms and also quantify the relative influence of the scene complexity measures on a given algorithm.

There is a growing trend of moving desktop applications to the Web by using AJAX to create user-friendly and interactive interfaces. Well-known examples include GMail, Hotmail, Google Wave and office applications. One common way to provide assurance about the correctness of such complex and evolving systems is through regression testing. Regression testing classical web applications has already been a notoriously daunting task because of the dynamism in web interfaces. AJAX applications pose an even greater challenge since the test case fragility degree is higher due to extensive run-time manipulation of the DOM tree, asynchronous client/server interactions, and (untyped) JAVASCRIPT. In this research, we propose a technique, in which we automatically infer a model of the web application and use this as an input for the test suite. We apply pipelined oracle comparators along with generated templates, to deal with the dynamic non-deterministic behavior in AJAX user interfaces. Our approach, implemented in the open source testing tool CRAWLJAX, provides a set of generic oracle comparators, template generators, and preconditions for handling the dynamic aspect of AJAX applications. We present a plugin that shows visualizations of test failure output in terms of differences between expected and observed DOM trees.We describe four case studies evaluating the effectiveness, scalability, and required manual effort of the approach and the applicability of CRAWLJAX in a real world production environment. The various improvements made to CRAWLJAX in this research are also discussed.

Continuous variable selection using shrinkage procedures have recently been considered as favorable models in a wide range of scientific research; in particular biomedical research. In some cases, it is desirable to select as few predictors as possible, to increase the interpretability of the attained prediction rule. One frequently used shrinkage procedure; the Lasso, imposes a L1 regularization on the regression coefficients of general linear models, inherently leading to sparse prediction rules. When dealing with multi-class prediction in generalized linear models each predictor has a regression coefficient for each class. A major disadvantage is that the Lasso selects individual regression coefficients instead of the more logical selection of predictors. In this paper, we demonstrate a new regularization procedure, based on the Group Lasso in multinomial logistic regression. This results in a lower number of retained predictors, but with similar prediction accuracy when compared to the regular Lasso regularization. To illustrate the new regularization applicability we have employed it on a large cohort of acute myeloid leukemia patients (AML, n=531) who are characterized on a gene expression microarray.

Process operators at large railway stations have the difficult task to secure a fluent train traffic flow while minimizing deviations from established timetables. Variation in actual train departure times is inevitable due to many circumstances such as arrival delays and fluctuations in alighting and boarding time, even if some buffer time is contained in the dwell time. Moreover, a departure may be delayed by waiting for a feeder train to secure a connection and by conflicting train movements prohibiting an outbound train path. The predictability of train processes is even more degraded when in similar situations different control actions are pursued depending on for instance individual dispatchers. In the Netherlands, passenger train services operate basically according to a cyclic timetable, repeating the same arrival and departure times each hour, with the exception of additional passenger trains in rush hours and freight trains that are scheduled in between the regular train services. It is hence anticipated that the traffic processes are mainly variations on a repetitious pattern. Analysis of historical realization data then yields operational insight that can be used to improve or support process management. To gain accurate operations data a software tool, TNV-Prepare, has been developed that filters relevant train detection data from train describer records. This paper starts with a brief account of the collection and preparation of train detection data. Then for the particular case of station Eindhoven a detailed punctuality analysis is reported including the performance of dwell and transfer times (tightness or possible recovery time) and train waiting times to secure connections. Departure delays are predicted from arrival delays using regression analysis, whereas the remaining noise is attributed to human factors.

We introduce here Cartesian splines or, for short, C-splines. C-splines are piecewise polynomials which are defined on adjacent Cartesian coordinate systems and are Cr continuous throughout. The Cr continuity is enforced by constraining the coefficients of the polynomial to lie in the null-space of some smoothness matrix H . The matrix-product of the null-space of the smoothness matrix H and the original polynomial base results in a new base, the so-called Cspline base, which automatically enforces Cr continuity throughout. In this article we give a derivation of this C-spline base as well as an algorithm to construct C-spline models.

Heterogeneous System Development needs Hardware/Software Partitioning performed early on in the development process. In order to do this early on predictions of hardware resource usage and delay are necessary. In this thesis a Quantitative Model is presented that can make early predictions to support the partitioning process. The model is based on Software Complexity Metrics, which capture important aspects of functions like control intensity, data intensity, code size, etc. In order to remedy the interdependence of the software metrics a Principal Component Analysis performed. The hardware characteristics were determined by automatically generating VHDL from C using the DWARV C-to-VHDL compiler. Using the results from the principal component analysis, the quantitative model was generated using linear regression. The error of the model differs per hardware characteristic. We show that for flip-flops the mean error for the predictions is 69%. In conclusion, our quantitative model can make fast and sufficiently accurate area predictions to support Hardware/Software Partitioning. In the future, the model can be extended by introducing extra software metrics, using more advanced modeling techniques, and using a larger collection of functions and algorithms.

A shift in consumer needs from predominantly functional to more psycho social needs, resulted in a new era within the Western economy: the Experience Economy. The ability to create experiences which keep speaking to ones imagination has become an indispensable ‘art’ of management, which is already being widely applied within the retail, merchandise and leisure sector. Within the market of moveable properties, the addition of an experience to a product or service, has already proven itself worthy. Project developers operating within the market of immovable properties seem to have noticed this as well, and started by integrating so called experience concepts within a number of housing projects. However, the added value of such an experience concept has never been demonstrated. This latter point is exactly where this research comes into play. By a semi-large survey amongst inhabitants of two (recent) Dutch housing projects, and by interviews of multiple project developers and real estate agents an attempt is made to get an insight in the added value an experience concept within a Dutch housing project, from a supply as well as from a demand side point of view.

In the online travel environment it is physically and economically not possible to retrieve the hotel room rates in real-time for every customer request. To overcome this problem the hotel rates are cached, but due to the fact that the suppliers will not send notifications of price changes it is a challenging task to keep the cache up to date. Analysis of the problem of how to improve the cache’s performance, in terms of accuracy and coverage, led to the conclusion that there is no single risk-free overall solution, but that the cache should be enhanced in a step-wise manner minimizing the (financial) impact of a faulty enhancement. The Enhancement Cycle is defined as a number of stages that each step-wise enhancement will go through. The system provides a solution for two of those stages. The system measures the cache performance in a correct manner and predicts the expiration time for a hotel rate. The system uses Complex Event Processing to detect the patterns needed to do the measurements. The system has been generalized and is extended with a Domain Specific Language for a low-effort application to other problems. The system is running in production and processes millions of messages a day. It aggregates and compresses the measurement data without loosing expressivity and its predictions improved the cache’s accuracy.

Aviation financers are interested in the current/future market value of used commercial aircraft, as this information is precious knowledge for them to support collateral position in the aircraft loan. In this paper, variables from general economy, airline industry and aviation fleet are explored to find out the factors predictive for used aircraft market. Two statistical methods- principal component regression and copula-are applied for building the prediction model of an aircraft.

This paper describes the efficiency of Dutch hospitals using the method of Data Envelopment Analysis (DEA). In particular the analysis focuses on explaining cost inefficiency measures due to each hospital’s operating environment. In previous works, the resulting DEA score is regressed on environmental factors via a Tobit approach. Previously, these approaches have been used (Simar and Wilson, J Prod Anal 7(1):63–80, 2000) but later these authors (Simar and Wilson 2007) demonstrated that bias is incurred since the efficiency score is a point estimate without a probability distribution around it that is required by the Tobit methodology. In this paper we use the Simar and Wilson bootstrapping techniques in order to obtain more efficient estimates of the environmental effects. It is shown that differences in estimated effects exist between the non-bootstrapped and bootstrapped models.

To understand the process of iron leaching and arsenic (As) mobilization, three bore wells were drilled in suspected iron and arsenic enriched areas of Jorhat, Assam, India, to study possible release and mobilization process in the aquifer. Sediments and groundwater samples, collected from different depths in these boreholes were analyzed for different parameters. Combined Eh-pH stability diagram of iron and arsenic indicated the presence of Fe (II) and As (III) species in the groundwater. BCR three step sequential extraction method indicated relative mobility of Fe and As in a similar trend of decreasing order as: Residual fraction > Reducible fraction > Exchangeable fraction > Oxidizable fraction. The digestion of aquifer sediments by D 3974-81, (ASTM) showed Fe (106 mg/kg to 26991 mg/kg), Mn (7 mg/kg to 1588 mg/kg), As (0.16 mg/kg to 18.6 mg/kg). The morphology and mineralogy of the aquifer sediments, studied by scanning electron microscope/energy dispersive x-ray (SEM/EDX) and x-ray diffraction (XRD) analysis indicated the levels of Fe and As concentrations in the aquifer sediments. Sequential heating in the muffle furnace indicated organic content (0.01 % to 13.7 %) and carbonate content (0.15 % to 9.19 %) of aquifer sediments. Laser particle size analyser was used to find the specific surface area of the sediments (0.02 m2/g to 0.19 m2/g). The lithology of the soil consisted of silt, very fine sand, fine sand, medium sand and fine gravel. The relationship between Fe, Mn, As, organic content, carbonate content and specific surface area of the sediments along the depth for different bore holes was interpreted separately including with the help of Principal Component Analysis (PCA). In multiple linear regression models, the experimentally found and predicted concentrations of iron correlated as 0.83 for B bore hole and 0.87 for E bore hole, whereas correlation coefficients for arsenic concentration were 0.88 in B bore hole and 0.80 in E bore hole. With the help of the linear regression models, the correlation between these elements and the factors controlling their concentrations were evaluated.

Introduction - The Dutch government wants to comply with European and global climate goals. One facet of this is reducing negative effects of car use by encouraging the use of cars that are more environmentally friendly, reducing car mileage, and increasing the use of more sustainable transport modes. Dutch policies make purchasing an environmentally friendly car more attractive, but they do not reduce the actual mileage. Therefore, these policies do not achieve the desired results of meeting the government’s environmental goals and reducing the negative effects of car usage. Company car policy - Company cars, on average, are newer and are more often equipped with diesel engines but they are also larger which makes the fuel efficiency worse than the efficiency of private cars. The use of company cars has developed from being a status symbol for board members and necessary for employees who have to travel frequently for work, to a common practice in the configuration of the salary package and are used as an incentive to attract talented people in specialized functions. The cost of a company car to the employee is fully compensated by the company regardless of his/her mileage, which could increase the private use of company cars. Employees only pay fiscal addition, which is a set amount each month. Therefore, people with a company car are insensitive to the marginal costs of car usage. With current policy, when people drive less than 500 km. privately, they do not have to pay fiscal addition. There are 260,000 people that use this option. New policy cuts this exemption and applies fiscal addition to all company cars. Commuter compensation policy - In case no company car is provided, commuter compensation can be provided by companies to cover an employee’s travel expenses to and from his/her work place. In the Netherlands, this benefit is exempt from tax up to 19 cents/km. Newly proposed policy suggests this compensation be no longer exempt from tax so that it will be added to the employee’s taxable income. Knowledge gaps - This research will focus on the following: comparing Dutch company car and commuter compensation policy to that of other countries (1); analyzing the effects of company cars and commuter compensation on mobility behavior (2); analyzing the effects of the mobility behavior of a household head on that of his/her partner and the other way around (3); analyzing the effects of company cars on the use of other modalities and substitution effects between those modalities (4); and analyzing what type of trips are affected by having a company car (5). Research overview - Based on literature research, the current and newly proposed Dutch policy regarding company cars and commuter compensation are compared to those of other countries. An overview of wanted and unwanted effects is presented per country, and best practices are selected. Furthermore, a conceptual model is constructed to form the theoretical basis to explain the findings from the explorative analyses. The analyses are done with multiple regression and structural equation modeling. We look at the effects of company cars and commuter compensation on mobility behavior in general, and on car usage in specific. This information is obtained by analyzing three datasets: one from the Netherlands in 1989, one from Germany in 2009, and one from the Netherlands in 2010. We look at people on a personal level, at the effects of men’s mobility behavior on that of women and the other way around, and finally at substitution effects between different modes of transport. Sub conclusion: policy comparison - The policy comparison indicates that both the current Dutch company car policy as the newly proposed company car policy incentivize people to purchase cleaner cars. The newly proposed policy affects all people, whereas current policy gives a break to people that stay below the threshold of 500 kilometers private usage of the company car. However, the marginal costs are still not paid by the company car users which remains a problem. This still means that people’s mobility behavior will not change dramatically. However, the negative effects of company car usage can be reduced because people are incentivized to take a cleaner car as company car. Combined with the taxation of commuter compensation, which should reduce the commuting mileage, the newly proposed Dutch policy seems to be more in line with the Dutch government’s goals of reducing emissions. By focusing the tax of commuter compensation on cars, the use of other modes can be stimulated. Sub conclusion: effects of having company cars on mobility behavior - The conclusion that can be drawn from analyzing all three datasets is that having a company car is related to more mobility. In the Dutch 1989 data we find that men with a company car make about 58% more car trips, and women with a company car about 44% more. If we look at generic data, which only has people with a job and a driver’s license in it, we find that having a company car is related to 70% more car trips. In addition, for this latter group, having a company car is related to 156% more commuting trips. In Germany we do not see a significant effect on the number of total trips, nor in the number of car trips if someone has a company car. We do, however, see more total travel time and total travel distance, as well as travel time by car and travel distance by car. Perhaps people in Germany use the company car to make longer trips. This could mean that people in Germany that have to travel further to work, or drive more for their work, get a company car. Most effects of company cars changed mobility behavior over a longer period of time prior to 2008. The Dutch data from 2010 also shows that having a company car is related to 26% more car trips and 45% more total travel time. Sub conclusion: effects of receiving commuter compensation on mobility behavior - When we look at the effects of commuter compensation in the generic Dutch data from 1989, we see that receiving compensation is related to 25% more commuting trips and 16% more car trips. A higher compensation amount is even related to 0.3% more car trips per extra euro of compensation. However, when we look at the household split data, we find that men receiving commuter compensation is related to 14% fewer total trips, and 83% fewer commuting trips for men. In addition, men receiving compensation is related to women making almost 5% fewer total trips per week. Sub conclusion: effects between men’s and women’s mobility behavior - The results from the Dutch 1989 data show that there were no significant effects from women’s mobility behavior on men’s mobility behavior. We do see that each trip that men make is related to women making 1.5% more trips. This was expected because men and women might make trips together. It is interesting that the effect does not appear the other way around. The German dataset, however, shows that there appears to be an enhanced effect between men and women regarding their generic mobility behavior. Each trip made by men is related to women making 1.1% more trips. Each trip made by women is related to men making 1.5% more trips. The German data shows that there is competition between men and women for car usage. Sub conclusion: other effects on mobility behavior - The most important other results are that the number of cars in a household and the education level have a positive correlation with mobility behavior. People with a higher education level are likely to have a higher income and thus more cars. These effects can probably be influenced by increasing car and/or road taxes because then every car becomes more expensive. Sub conclusion: substitution effects between modalities - If we look at Dutch and German results for substitution effects between different modalities, we find that company cars in The Netherlands are related to a higher number of car trips and a lower number of trips by all other modalities other than walking. In Germany there are no significant effects of having a company car. This confirms our expectations that a company car makes it less appealing to take other modes of transport than the car in The Netherlands. It is interesting to see that this same effect is not found in Germany. The fact that walking trips are not affected by having a company car makes sense, because sometimes taking the car for short distances is irrational. For both countries, we see that all other modalities form a substitution for car trips, and the other way around. The number of car trips increases is negatively correlated to the number of trips by all other modes. Furthermore, the number of train trips is negatively correlated to the number of car trips in both data sets, and that the Dutch 2010 data shows that bicycle trips are also a substitute for train trips. The number of bicycle trips has a negative correlation to the number of car and bus trips in both data sets. However, the Dutch 2010 data also show that train trips are a substitute for bicycle trips. Perhaps the bicycle is a suitable alternative for the train in The Netherlands, but not so much in Germany. Finally, car trips are the only substitute for walking. Sub conclusion: purpose of trips - Having a company car is positively related to the number of trips for work, and the number of transport-related trips. Having a company car is related to a lower number of shopping trips. Policy implications - The fact that proposals are being written to reduce the negative effects of cars is a good thing. In The Netherlands, current and newly proposed company car policies stimulate buying cleaner cars, but they need to incorporate an incentive for the user to reduce car mileage. A personal contribution to the marginal costs seems like a good way to do this. This makes company car users more aware of the costs and it is a financial incentive for them to reduce their mileage. The commuter compensation taxation should reduce mobility behavior. However, the tax is planned to also be levied on public transport modalities. By focusing the tax on cars, car mileage can be reduced because people would have a financial incentive to travel with other modes of transport. Making it less attractive to own multiple cars is also something that could reduce mileage. This can be done by increasing the road and/or car taxes but also in a new way like taxing second cars more than first cars, as is done with houses. The fact that the new policy is dismissed in the fall of 2012 is hopefully but a bump in the road for the way to a more sustainable future of road transport.

We present a computer-aided detection (CAD) system for computed tomography colonography that orders the polyps according to clinical relevance. TheCADsystem consists of two steps: candidate detection and supervised classification. The characteristics of the detection step lead to specific choices for the classification system. The candidates are ordered by a linear logistic classifier (logistic regression) based on only three features: the protrusion of the colon wall, the mean internal intensity, and a feature to discard detections on the rectal enema tube. This classifier can cope with a small number of polyps available for training, a large imbalance between polyps and non-polyp candidates, a truncated feature space, unbalanced and unknown misclassification costs, and an exponential distribution with respect to candidate size in feature space. Our CAD system was evaluated with data sets from four different medical centers. For polyps larger than or equal to 6mmwe achieved sensitivities of respectively 95%, 85%, 85%, and 100% with 5, 4, 5, and 6 false positives per scan over 86, 48, 141, and 32 patients. A cross-center evaluation in which the system is trained and tested with data from different sources showed that the trained CAD system generalizes to data from different medical centers and with different patient preparations. This is essential to application in large-scale screening for colorectal polyps.

We study nonparametric estimation of convexregression and density functions by methods of least squares (in the regression and density cases) and maximum likelihood (in the density estimation case).We provide characterizations of these estimators, prove that they are consistent and establish their asymptotic distributions at a fixed point of positive curvature of the functions estimated. The asymptotic distribution theory relies on the existence of an “invelope function” for integrated two-sided Brownian motion $+t^4$ which is established in a companion paper by Groeneboom, Jongbloed and Wellner.