This thesis report details a project to extend the AdGateway product developed at Dohi Sweden with a framework that allows advertisers to create interactive advertisements. The key feature of this framework is to provide augmented reality capabilities by taking advantage of face recognition and webcams on smart TV systems. The interactive advertisements are created as applications in JavaScripts which get downloaded on the fly when a user clicks on a banner for that type of advertisement. This implementation will both extend the back-end system to support the distribution of this new type of advertisement as well as the front-end for the supported devices in order to set up the environment and run the JavaScript application. The Sony Google TV is the primary focus for the implementation but an extensive study is also included in the report to gather information about other possible platforms that could be supported. This report also studies in detail how the performance of face recognition can be improved by performing illumination normalization on input images. The project could not be completed on time due to limitations in both hardware and software regarding face detection. Nevertheless a framework has been implemented, and solutions for the impediments are discussed in the report

Ad Gateway is a product which provides targeting of online advertisements. This thesis focuses on improving its web based admin interface, used for defining targeting properties as well as creating the advertisements. First by integrating the system with third party web services used for storage of advertisement campaigns, called advertisement providers, allowing advertisement to be retrieved and uploaded directly to their servers. Later to extend the admin interface with new functionality, including the ability preview advertisement campaigns before pushing them to a live environment. The thesis performs an evaluation of different technologies for developing rich internet applications, such as HTML5, JavaScript, Adobe Flash and Microsoft Silverlight. HTML5 and JavaScript was determined to be best suited technologies for implementing the new functionality.

In this Master's project increasing the performance of object detection in forestry environment, based on the extracted features is studied. There are several object detection projects for robots which are based on feature calculation and extraction. An example of these kinds of projects is the sugarbeet project [3] that has inspired the feature selection and calculations parts presented in this report. Extracted feature sets are given to several classifiers and their results are merged and fused such that the overall performance of the forestry object detection increases. Furthermore different supervised and unsupervised methods of dimensionality reduction are applied on the feature set as an approach to improve classiffcation accuracy. Comparison between the output classification performance of dimensionality reduction methods show that applying supervised methods result in improving the classification performance by about 12 percent.

The project involves implementing a finished dispersion model into ArcGIS. The goal of the tool is to show how dangerous and toxic substances will travel in the ground after long periods. The goal of the project is to understand GIS in general, what it is used for and gain an insight into how developing tools for ArcGIS is, what challenges might exists. Understanding the type of data that can be stored and accessed in ArcGIS a long with the tools and functionality offered by the system when using Python as the main language for development. A prototype of the tool is created.

When visualising a natural phenomena, such as a gas cloud, the constant movement and turbulens can pose a complex problem. With the use of a stochastic particle dispersion model a gas cloud is simulated and stored in a datafile. This thesis report describes some possible methods on how to structure and render such a datafile in an effective and realistic manner. This is achieved by using interpolation, the octree data structure and volume rendering. The end product will be used in a larger project at FOI, with the goal to facilitate the training of personal. Three different rendering methods are implemented and compared against each other. The best one will build the foundation for the FOI project. The end result of this report shows that the methods could be used with satisfying results but more development is needed.

There is a gap in the understanding of software quality between developers and non-technical stakeholders, the software quality gap, which leads to disagreements about the amount of time that should be used for quality improvements. The technical debt metaphor reduces this gap some by describing software quality in economic terms, enabling developers and non-technical stakeholders to communicate about the quality. However, the metaphor is vague and not very concrete in explaining the gap. The purpose of this thesis is to concretize the technical debt metaphor using Domain-Driven Design, an approach in which communicating software characteristics is central, in order to reduce the software quality gap further. Using the terminology of Domain-Driven Design, a new concept is defined: model debt, which measures and communicates the software quality of the domain model. An application is built, the ModelDebtOMeter, which extracts the domain model of a software system and visualizes it along with its corresponding model debt. The model debt of a legacy system is amortized and the results of the amortizations are presented to the non-technical stakeholders of the system using the ModelDebtOMeter, allowing the non-technical stakeholders to evaluate the model debt concept. The results show that the non-technical stakeholders think that model debt is an understandable concept which reduces, but not eliminates, the software quality gap.

In our daily life, we are faced with many occluded faces. The occlusion may be from different objects like sunglasses, mufflers, masks, scarves etc. Sometimes, this occlusion is used by the criminal persons to hide their identity from the surroundings. In this thesis, a technique is used to detect the facial occlusion automatically. After detecting the occluded areas, a method for image reconstruction called aPCA (asymmetrical Principal Component Analysis) is used to reconstruct the faces. The entire face is reconstructed using the non occluded area of the face. A database of images of different persons is organized which is used in the process of reconstruction of the occluded images. Experiments were performed to examine the effect of the granularity of the occlusion on the aPCA reconstruction process. The input mask image is divided into different parts, the occlusion for each part is marked and aPCA is applied to reconstruct the faces. This process of image reconstruction takes a lot of processing time so pre-defined eigenspaces are introduced that takes very less processing time with very less quality loss of the reconstructed faces.

In [Dre09], Drewes outlines Marbles, a programming framework for working in a generic and systematic way, not only on trees, as several frameworks already exist for this purpose, but on tree recognisers, transducers, generators and other formal devices as well. This thesis presents a prototype of a proposed implementation of this framework, demontrating its functionality by using it as a base for implementing a well-known algorithm on tree transducers.

Automatic classification of scientific articles based on common characteristics is an interesting problem with many applications in digital library and information retrieval systems. Properly organized articles can be useful for automatic generation of taxonomies in scientific writings, textual summarization, efficient information retrieval etc. Generating article bundles from a large number of input articles, based on the associated features of the articles is tedious and computationally expensive task. In this report we propose an automatic two-step approach for topic extraction and bundling of related articles from a set of scientific articles in real-time. For topic extraction, we make use of Latent Dirichlet Allocation (LDA) topic modeling techniques and for bundling, we make use of hierarchical agglomerative clustering techniques. We run experiments to validate our bundling semantics and compare it with existing models in use. We make use of an online crowdsourcing marketplace provided by Amazon called Amazon Mechanical Turk to carry out experiments. We explain our experimental setup and empirical results in detail and show that our method is advantageous over existing ones.

In this study, we propose textual summarization for scientific publications and mobile phone usage patterns. Textual summarization is a process that takes a source document or set of related documents, identifying the most salient information and conveying it in less space than the original text. The increasing availability of information has necessitated deep research for textual summarization within Information Retrieval and the Natural Language Processing (NLP) area because textual summaries are easier to read, and provide to access to large repositories of content data in an efficient way. For example, snippets in web search are helpful for users as textual summaries. While there exists summarization tools for textual summarization, either they are not adapted to scientific collection of documents or they summarize short form of text such as news. In the first part of this study, we adapt the MEAD 3.11 summarization tool [19] to propose a method for building summaries of a set of related scientific articles by exploiting the structure of scientific publications in order to focus on some parts that are known to be the most informative in such documents. In the second part, we generate a natural language statement that describes a more readable form of a given symbolic pattern extracted from Nokia Challenge data. The reason is that the availability of mobile phone usage details enables new opportunities to provide a better understanding of the interest of user populations in mobile phone applications. For evaluating the first part of study, we make use of Amazon Mechanical Turk (Mturk) to validate summarization output.