https://www.um.edu.mt/library/oar/handle/123456789/21921 2026-06-14T12:02:14Z 2026-06-14T12:02:14Z https://www.um.edu.mt/library/oar/handle/123456789/73653 2021-04-14T13:44:31Z 2017-01-01T00:00:00Z

Title: Application and improvement of genetic algorithms and genetic programming towards the fight against spam and other internet malware Abstract: While spam is increasingly acknowledged as a very expensive problem on the Internet, spam filters attempt to detect spam from emails with the highest possible accuracy. Reverse Polish Notation (RPN) expressions are proposed as a means to combine a range of evaluated features from emails for the detection of spam. Theoretical arguments for the use of RPN expressions applied to spam detection are proposed, together with a new RPN Block Representation. Theoretical comparisons of RPN expressions with Naïve Bayes and Support Vector Machine are also given. It is shown that such RPN expressions are more expressive. A proof that email spam detection is NP-complete is given by mapping groups of email spams onto malware virus families. Seventy-two features, ranging from Subject-line, Header-based, Message Body-based, URL-based and stylistic, have been used to evolve RPN expressions using Linear Genetic Programming. New features and specifically the application of a group of URL features to spam detection are proposed (since many spams contain links to domains which are at times even malicious). These new features are shown to be useful for spam detection theoretically and in practice. Linear Genetic Programming is a subset of Genetic Programming where chromosomes are computer programs represented using imperative computer language instructions or machine code instead of trees made up of symbolic expressions. Such machine code can encode RPN expressions. The Linear Genetic Programming system is used to “learn” an RPN expression consisting of a combination of features which can be used to detect spam. A number of feature selection algorithms are used to identify which subsets of features are most relevant to classification. The feature selection techniques Minimum Redundancy Maximum Relevance method (this filter technique finds features which are mutually far from each other while still having high “correlation” to classification), the conventional Maximum Relevance method, Principle Component Analysis and using the entire feature set are investigated using the Linear Genetic Programming system applied to the SpamAssassin Spam Corpus, which is a standard ham and spam archive. Theoretical and practical comparisons with literature results and industry open source applications are given for the proposed system. A new Block Crossover operator which helps preserve and keep intact RPN Blocks as well as a new RPN First Subexpression algorithm are also proposed. This algorithm helps to maintain a double check for spam and ham and results in improvement of detection accuracy and other metrics. The system was very demanding computationally, taking a long time to run on a supercomputer. The proposed system achieved 99.17% accuracy and an F1-score of 0.9868, which compare very well with results given in the literature as well as with performance of industrial Bogofilter and SpamAssassin Spam Filters. Description: PH.D.

2017-01-01T00:00:00Z https://www.um.edu.mt/library/oar/handle/123456789/41423 2019-03-21T02:21:15Z 2017-01-01T00:00:00Z

Title: Indoor location system Abstract: The objective of this project is to design and develop a mobile oriented, indoor location technique. Existing outdoor triangulation positioning techniques like Global Positioning System (GPS) and Globalnaya Navigazionnaya Sputnikovaya Sistema (GLONASS) prove futile and are generally not suitable or accurate enough for indoors, particularly for multilevel buildings. The set goal is to design a system that enables a single Wi-Fi access point (AP) to locate the client device, without support from any additional infrastructure in place. Further, no prior radio-frequency finger-printing procedure is required, thus reducing the computational complexity significantly. This report reviews attempts to design a novel positioning approach with the help of current knowledge. The method suggested in this report is a hybrid angle of arrival (AoA) and received signal strength (RSS) technique. By making use of ray tracing to characterise the propagation in a multi-path rich environment and taking advantage of beamforming and the multiple-input and multiple-output (MIMO) system, a fairly accurate estimation can be obtained. Making use of a ray-tracing analysis, as presented herein, allows for site-specific location positioning using only a single AP. The objective of the report is to present the methodologies and complications encountered in implementing such a system and to present a preliminary analysis of calculations, measurement accuracy and simulation of the suggested algorithm. Description: B.SC.(HONS)COMPUTER ENG.

2017-01-01T00:00:00Z https://www.um.edu.mt/library/oar/handle/123456789/38983 2019-02-06T10:30:58Z 2017-01-01T00:00:00Z

Title: Hand drawn sketch drawings to vector graphics Abstract: Computer Numerical Control (CNC) machines and 3D printers are becoming more accessible. This allows artists and consumers to create physical objects from their drawn ideas. However, many artists are more familiar with the traditional method of sketching which uses raster format. For these kind of devices, a special instruction set based on vector notations is used which requires specialised software. One problem is that learning a new interface or technology for 3D modelling is not trivial and some might give up before trying. A bridge that easily connects these two worlds would be beneficial to both. However, even though the interpretation of sketches appears to be trivial for humans, so cannot be said for machines. This holds particularly true when artists use artistic cues such as shadows to represent depth. Additionally, hand drawn sketches are intrinsically imperfect and might contain curves making the gap between raster to vector hard to reduce. In this research, a method that automatically converts hand-drawn sketches in presence of shadows and curves is presented. The following study is divided into two section. The first section deals with junction localisation and identification to ensure that the topological fidelity of the drawing is retained. When compared to current state of the art, the results obtained shows an improvement of 61% when the proposed methodology was evaulated for junction spatial localisation using Salient Point Error over the same dataset. Even though junction type identification was not used during the proposed vectorisation pipeline, a number of methods were described and evaluated for junction type classification. Classification was performed using three methodologies and the best classification results obtained an F-score of 0.95. The second section dealt with contour extraction to remove shadows and other artefacts from the drawings. Each pixel was assigned an orientation based on the direction of the surrounding pixels and used the result was used to identify the path between two connected junctions. Unconnected lines and recovery of missed junctions were also considered. An average F-measure score of 0.992 was obtained over the whole dataset when the ground truths and the vectorised images were compared using a contour evaluation protocol. A dataset of 17 images was used to cover drawings created in sketching software and on paper using straight and curved lines. The reconstruction was performed using either lines, arcs or splines as deemed the most adequate. We showed that our method performed better in junction and contour detection and the results obtained were consistent throughout the whole dataset including straight and curved lines drawings. Description: M.SC.ARTIFICIAL INTELLIGENCE

2017-01-01T00:00:00Z https://www.um.edu.mt/library/oar/handle/123456789/38859 2019-02-12T08:37:23Z 2018-01-01T00:00:00Z

Title: Semi-automatic segmentation of human anatomical imagery Abstract: Manual segmentation of anatomical imagery is a challenging and laborious task which this dissertation attempts to alleviate. We present a semi-automatic segmentation system which operates on a new data set of photographic human anatomical imagery. A morphological tree-based segmentation method was utilised in order to reach this aim. We placed a particular focus on elongated structures in order to demonstrate the e ectiveness of the algorithms. The resultant outputs were presented to academics in the anatomical sciences for evaluation. Qualitative and quantitative results which were collected throughout the course of the experimentation phase indicate that the system was successful in producing meaningful labelled segmentation outputs with particularly good performance on elongation, which were commended by the experts. We believe that these results provide a good initialisation step for more re ned labelled images which can be used in a number of di erent professional and educational tools. Furthermore, the outcome of this dissertation demonstrates that a technical window exists in this area, and a foundation for further research has been created in this work. Description: M.SC.ARTIFICIAL INTELLIGENCE

2018-01-01T00:00:00Z