Publications
see more in google scholar
2025
- IEEE BIBEFusion Strategies for Breast Cancer Characterization Using Traditional and Deep Learning ModelsPedro Lima, Jaime Cardoso, and Hélder OliveiraIn Proceedings of the 25th International Conference on BioInformatics and BioEngineering, 2025
Breast cancer remains one of the most prevalent and deadly cancers worldwide, making accurate evaluation of molecular markers important for effective disease management. Biomarkers such as ER, PR, and HER2 are typically assessed because they help inform prognosis and guide treatment decisions. Predicting these characteristics from imaging can support earlier clinical intervention, reduce reliance on invasive procedures, and contribute to more personalized care. While radiomics and deep learning approaches have demonstrated potential, comprehensive comparisons across these methods are still limited. This study evaluated handcrafted features, deep features, and end-to-end deep learning models for predicting ER, PR, and HER2 status from DCE-MRI. Each feature type was first assessed individually and then combined using early and late fusion. Handcrafted and deep features were processed through a pipeline that included resampling, dimensionality reduction, and model selection, while end-to-end models were trained using different initialization strategies and loss functions. The best models achieved AUCs of 0.659 for ER, 0.679 for PR, and 0.686 for HER2. Although late fusion generally improved performance, bias toward the majority classes persisted. Overall, the results suggest that combining different modeling strategies may enhance robustness in breast cancer characterization.
2024
- CRC PressSkin Cancer and Hansen’s Disease DiagnosisPedro Lima, Juliana Gomes, Laís Castro, and 3 more authorsIn Biomedical Imaging, 2024
The advancement of the use of Artificial Intelligence (AI) in the healthcare sector makes it possible to use computational intelligence applications to assist healthcare professionals in the diagnosis process, facilitating and optimizing early detection and allowing for a more accurate diagnosis. The application of machine learning methods, and, more recently, deep learning, has shown promising results. These approaches allow powerful tools to support diagnostic imaging and signs to be built, through the extraction of image features and the creation of a classification system, for example. There are several diseases known and classified by man, with different causes and prevalence. Therefore, contributing to the early detection of diseases defined as neglected was the initial motivation for this work.
2023
- Zero-Phase FIR Filter Design Algorithm for Repetitive ControllersPedro Lima, Rafael Neto, Francisco Neves, and 3 more authorsEnergies, 2023
Repetitive controllers (RCs) are linear control structures based on the internal model principle. This control strategy is known for its ability to control periodic reference signals, even if these signals have many harmonic components. Despite being a solution that results in a good performance, several parameters of the repetitive controller need to be correctly tuned to guarantee its stability. Among these parameters, one that has high impact on the system performance and stability is the finite impulse response (FIR) filter, which is usually used to increase the stability domain of RC-based controllers. In this context, this paper presents a complete tutorial for designing the zero-phase FIR filter, which is often used to stabilize control systems that use RC-based controllers. In addition, this paper presents a Matlab® application developed for performing the stability analysis of RC systems and designing its FIR filter. Simulation and experimental results of a shunt active power filter are used to validate the algorithm and the Matlab® application.
2022
- CRC PressClinical Decision Support in the Care of Symptomatic Patients with COVID-19: An Approach Based on Machine Learning and Swarm IntelligenceIngrid Nunes, Pedro Lima, Andressa Ribeiro, and 10 more authorsIn Swarm Intelligence, 2022
In December 2019, the city of Wuhan, China became the scene of an outbreak of pneumonia without a defined cause or etiologic agent. The event attracted attention beyond the Chinese government, such as the World Health Organization (WHO). Patients infected or suspected of contamination were isolated, and people in contact with them were monitored. Due to its high contamination rate, COVID-19 is considered one of the greatest sanitary challenges on a global scale of the century. Technology is a strong ally in helping healthcare services. In the context of the pandemic, an intelligent system can be of great help to support medical decisions, aiding in decision-making regarding the patient’s hospitalization in order to optimize the prognosis. This work aims to evaluate the performance of intelligent classifiers, to be used to support decision making regarding the type of care that the patient should receive. Patient data correspond to hematological and biochemical tests. Four families of smart classifiers were investigated. The most relevant hematological parameters for decision making were also analyzed using the particle swarm optimization algorithm. The results showed that it is possible to build intelligent decision support systems aimed at attending Covid-19 with good results (accuracies above 80%) based on the most relevant hematological parameters.
2019
- IEEE SVR3DJPi: An Open-Source Web-Based 3D Simulator for Pololu’s 3Pi PlatformLucas Maggi, João Teixeira, José Júnior, and 4 more authorsIn Proceedings of the 21st Symposium on Virtual and Augmented Reality, 2019
Line-following robots can recognize and follow a line drawn on a surface. Their operating principles have elements that could be used in the development of numerous autonomous technologies, with applications in education and industry. This class of robots usually represent the first contact students have with educational robotics, being used to develop students’ logic thinking and programming skills. The cost of robotic platforms is still prohibitive in low-budget schools and universities, which makes almost impossible having a platform for each small group of students in a classroom, harming the learning process. This work proposes a 3D web-based open-source simulator for Pololu’s 3Pi line-following robots, making such technology more accessible and available even for distance learning courses. The developed software simulates the robot’s physical structure, behavior, and operations - as being able to read surfaces -, enabling the user to observe the robot following the line as the code commands. The simulator was validated based on experiments that included motion analysis and time measurements of pre-stablished tasks so that its execution could be more coherently based on what happens in reality.
2018
- IEEE LARSImproving Early Robotics Education Using a Line-Following Robot SimulatorPedro Lima, Maria Bezerra, Ana Tavares, and 5 more authorsIn Proceedings of the 15th Latin American Robotics Symposium, 2018
Line-following robots have the ability to recognize and follow a line drawn on a surface. Elements of their operating principles could be used in the evelopment of numerous autonomous technologies, with applications in education and industry. A simulator has been developed to aide in performing several trials in order to validate a project. By taking the Pololu 3pi Robot as the model, the proposed solution simulates its physical structure, behavior, and operations - being able to read lines on surfaces - enabling the user to observe the robot following the line according to the code used. This paper aims to validate the developed simulator as an alternative to ease the process of learning to use the 3pi platform applied in both educational and competitive environments.
- IEEE LARSTurning Pololu 3Pi into a Multi-Programming PlatformJosé Júnior, Pedro Lima, Maria Bezerra, and 2 more authorsIn Proceedings of the 15th Latin American Robotics Symposium, 2018
Line-following robots have the ability to recognize and follow a line drawn on a surface. It works based on a simple self-sustainable system composed with a set of sensors, motors and a controller. In order to get optimal performance in such robots, it is necessary to carry out several tests to evaluate the behavior in each trial. In the majority of cases, a new trial requires to upload a new program, thus slowing down the development of the line-following. This paper presents an approach to solve the inconvenience of having to upload a new program in each trial. It consists in merging multiple codes in to one to create a program that gives the user the ability to switch between them anytime inside Pololu’s 3pi line follower platform.
- SBASimulador de Código Aberto para a Plataforma Pololu 3PiJosé Júnior, Maria Bezerra, Pedro Lima, and 3 more authorsIn Proceedings of the 22nd Congresso Brasileiro de Automática, 2018
Robôs seguidores de linha têm a capacidade de detectar e seguir uma linha desenhada em uma superfície. Possuem em seu principio de funcionamento elementos que servem como referência para o desenvolvimento de várias outras tecnologias autônomas, com aplicações em educação e na indústria. O simulador desenvolvido visa facilitar uma das problemáticas que envolvem este tipo de ferramenta: a necessidade de realizar diversos testes para validar um projeto. Usando como modelo físico o Pololu 3pi Robot, o software simula seu funcionamento, age de acordo com a estrutura do robô real e é capaz de realizar leituras da superfície, permitindo observar o desempenho do robô ao seguir a linha operando com determinado código. O simulador desenvolvido foi validado a partir de experimentos que envolveram análise de movimentos e medições de tempo de execução de tarefas previamente estabelecidas, para que sua resposta fosse análoga ao que acontece na realidade.