ESTG - Capítulos de livros
URI permanente para esta coleção:
Navegar
Percorrer ESTG - Capítulos de livros por Domínios Científicos e Tecnológicos (FOS) "Ciências Naturais::Ciências da Computação e da Informação"
A mostrar 1 - 10 de 18
Resultados por página
Opções de ordenação
- 3D shape prior active contours for an automatic segmentation of a patient specific femur from a CT scanPublication . Almeida, D.; Folgado, J.; Fernandes, P.R.; Ruben, RuiThe following paper describes a novel approach to a medical image segmentation problem. The fully automated computational procedure receives as input images from CT scan exams of the human femur and returns a three dimensional representation of the bone. This patient specific iterative approach is based in 3D active contours without edges, implemented over a level set framework, on which the evolution of the contour depends on local image parameters which can easily be defined by the user but also on a priori information about the volume to segment. This joint approach will lead to an optimal solution convergence of the iterative method. The resulting point cloud can be an excellent starting point for a Finite Element mesh generation and analysis or the basis for a stereolitography for example.
- Adaptive Evolutionary Testing: An Adaptive Approach to Search-Based Test Case Generation for Object-Oriented SoftwarePublication . Ribeiro, José Carlos Bregieiro; Zenha-Rela, Mário Alberto; de Vega, Francisco FernándezAdaptive Evolutionary Algorithms are distinguished by their dynamic manipulation of selected parameters during the course of evolving a problem solution; they have an advantage over their static counterparts in that they are more reactive to the unanticipated particulars of the problem. This paper proposes an adaptive strategy for enhancing Genetic Programming-based approaches to automatic test case generation. The main contribution of this study is that of proposing an Adaptive Evolutionary Testing methodology for promoting the introduction of relevant instructions into the generated test cases by means of mutation; the instructions from which the algorithm can choose are ranked, with their rankings being updated every generation in accordance to the feedback obtained from the individuals evaluated in the preceding generation. The experimental studies developed show that the adaptive strategy proposed improves the test case generation algorithm's efficiency considerably, while introducing a negligible computational overhead.
- Analysis of Power Quality Disturbances in Industry in the Centre Region of PortugalPublication . Moreira, Licinio; Leitão, Sérgio; Vale, Zita; Galvão, João; Franco Marques, Pedro JoséPower quality issues have taken a more prominent role in power systems over the last years. These issues are of major concern for energy customers, primarily for customers with a widespread use of electronic devices in their manufacturing processes. Even though the quality of service is increasing, customers are becoming more demanding of the energy provider. This research aims to provide some industrial managers the technical support in deciding of investments in the mitigation of power quality disturbances, such as the use of less sensitive devices or the use of interface devices (UPS, DVR...) In order to recommend an appropriate solution, the problem is characterized. The technical and economic influences of the PQ disturbances in the manufacturing processes are assessed resorting to power quality audits in the customer facilities. This research covered a significant number of facilities in several industrial activities.
- Applying Scatter Search to the Location Areas ProblemPublication . Luz, Sónia Maria Almeida da; Vega-Rodríguez, Miguel A.; Gómez-Pulido, Juan A.; Sánchez-Pérez, Juan M.The Location Areas scheme is one of the most common strategies to solve the location management problem, which corresponds to the management of the mobile network configuration with the objective of minimizing the involved costs. This paper presents a new approach that uses a Scatter Search based algorithm applied to the Location Areas scheme as a cost optimization problem. With this work we pretend to analyze and set the main parameters of scatter search, using four distinct test networks and compare our results with those achieved by other authors. This is a new approach to this problem and the results obtained are very encouraging because they show that the proposed technique outperforms the existing methods in the literature.
- A Face Attention Technique for a Robot Able to Interpret Facial ExpressionsPublication . Simplício, Carlos; Prado, José; Dias, JorgeAutomatic facial expressions recognition using vision is an important subject towards human-robot interaction. Here is proposed a human face focus of attention technique and a facial expressions classifier (a Dynamic Bayesian Network) to incorporate in an autonomous mobile agent whose hardware is composed by a robotic platform and a robotic head. The focus of attention technique is based on the symmetry presented by human faces. By using the output of this module the autonomous agent keeps always targeting the human face frontally. In order to accomplish this, the robot platform performs an arc centered at the human; thus the robotic head, when necessary, moves synchronized. In the proposed probabilistic classifier the information is propagated, from the previous instant, in a lower level of the network, to the current instant. Moreover, to recognize facial expressions are used not only positive evidences but also negative.
- FlexSPMF: A Framework for Modelling and Learning Flexibility in Software ProcessesPublication . Martinho, Ricardo; Varajão, João; Domingos, DulceSoftware processes are dynamic entities that are often changed and evolved by skillful knowledge workers such as software development team members. Consequently, flexibility is one of the most important features within software process representations and related tools. However, in the everyday practice, team members do not wish for total flexibility. They rather prefer to learn about and follow previously defined advices on which, where and how they can change/adapt process representations. In this paper we present FlexSPMF: a framework for modelling controlled flexibility in software processes. It comprises three main contributions: 1) identifying a core set of flexibility concepts; 2) extending a Process Modelling Language (PML)'s metamodel with these concepts; and 3) providing modelling resources to this extended PML. This enables process engineers to define and publish software process models with additional (textual/graphical) flexibility information. Other team members can then visualise and learn about this information, and change processes accordingly.
- Fraud Prediction in Smart Supply Chains Using Machine Learning TechniquesPublication . Constante-Nicolalde, Fabián-Vinicio; Guerra-Terán, Paulo; Pérez-Medina, Jorge-LuisIn the domain of Big Data, the company’s supply chain has a very high-risk exposure and this must be observed from a preventive perspective, that is, act before such situations occur. As a company grows and diversifies the number of suppliers, customers and therefore increases its number of daily transactions and associated risks. Despite the innovation and improvements that have been incorporated into financial management, credit and debit cards are the main means of exchanging cash online, with the expansion of e-commerce, online shopping has also increased number of extortion cases that have been identified and that continues to expand greatly. It takes a lot of time, effort and investment to restore the impact of these damages. In this paper, we work with machine learning techniques, used in predicting smart supply chain fraud, are valuable for estimating, classifying whether a transaction is normal or fraudulent, and mitigating future dangers.
- A Hybrid Differential Evolution Algorithm for Solving the Terminal Assignment ProblemPublication . Bernardino, Eugénia; Bernardino, Anabela; Sánchez-Pérez, Juan Manuel; Gómez-Pulido, Juan Antonio; Vega-Rodríguez, Miguel AngelThe field of communication networks has witnessed tremendous growth in recent years resulting in a large variety of combinatorial optimization problems in the design and in the management of communication networks. One of these problems is the terminal assignment problem. The task here is to assign a given set of terminals to a given set of concentrators. In this paper, we propose a Hybrid Differential Evolution Algorithm to solve the terminal assignment problem. We compare our results with the results obtained by the classical Genetic Algorithm and the Tabu Search Algorithm, widely used in literature.
- Improving Text Classification Performance with Incremental Background KnowledgePublication . Silva, Catarina; Ribeiro, BernardeteText classification is generally the process of extracting interesting and non-trivial information and knowledge from text. One of the main problems with text classification systems is the lack of labeled data, as well as the cost of labeling unlabeled data. Thus, there is a growing interest in exploring the use of unlabeled data as a way to improve classification performance in text classification. The ready availability of this kind of data in most applications makes it an appealing source of information. In this work we propose an Incremental Background Knowledge (IBK) technique to introduce unlabeled data into the training set by expanding it using initial classifiers to deliver oracle decisions. The defined incremental SVM margin-based method was tested in the Reuters-21578 benchmark showing promising results.
- Improving Visualization, Scalability and Performance of Multiclass Problems with SVM Manifold LearningPublication . Silva, Catarina; Bernardete RibeiroWe propose a learning framework to address multiclass challenges, namely visualization, scalability and performance. We focus on supervised problems by presenting an approach that uses prior information about training labels, manifold learning and support vector machines (SVMs). We employ manifold learning as a feature reduction step, nonlinearly embedding data in a low dimensional space using Isomap (Isometric Mapping), enhancing geometric characteristics and preserving the geodesic distance within the manifold. Structured SVMs are used in a multiclass setting with benefits for final multiclass classification in this reduced space. Results on a text classification toy example and on ISOLET, an isolated letter speech recognition problem, demonstrate the remarkable visualization capabilities of the method for multiclass problems in the severely reduced space, whilst improving SVMs baseline performance.