Portal de Periódicos da Univali (Universidade do Vale do Itajaí)
Not a member yet
9513 research outputs found
Sort by
Uso de um sistema de vídeo para estudo ecológico do macro e megazooplâncton em águas costeiras
Get PDFUnderwater images obtained using digital cameras have become one of the main tools for studying the behavior and ecology of marine organisms in their native environment. In this research, the acquisition of images by trawling and vertical profiling with a commercial camera mounted on a steel structure called a Bell was tested. Fifty-four videos of longer horizontal trawls (stratified) (10 minutes) and 27 videos of short vertical profiling (3 minutes) were obtained, and the quality of the images associated with the identification of organisms was evaluated. The best way to work with the Bell, the need for accessories and adaptations and the protocol for its use were determined, as well as the relationship between water transparency (Secchi disk) and the possibility of identifying and quantifying planktonic forms in the images. The results indicate that the use of the Bell at low speed, i.e., in vertical profiling and with lighting, presented the best results and greater efficiency in identifying the occurrences of organisms. The relationship between water transparency and obtaining good quality images was observed more reliably in conditions of Secchi disk measurements greater than 2.5 m. Due to the water transparency conditions and the variability in the occurrence of organisms, the probability of recording gelatinous forms was 37.5% of the total field trips. Analysis of the videos indicated a possible vertical migration behavior of the Hydrozoa Olindias sambaquiensis. The images also highlighted the abundance of Pseudo-TEPs (Transparent Exopolymer Particles), structures believed to play a relevant role in the pelagic ecology of coastal waters.Imagens subaquáticas obtidas com o uso de câmeras digitais tem se tornado uma das principais ferramentas para estudos, no próprio ambiente, do comportamento e da ecologia de organismos marinhos. Com esse objetivo testou-se a obtenção de imagens por arrasto e por perfil vertical de uma câmera comercial montada em uma estrutura de aço denominada de Sino. Foram obtidas 54 vídeos de arrastos horizontais (estratificadas) de maior duração (10 minutos) e 27 vídeos de perfilagem vertical de curta duração (3 minutos) e avaliado a qualidade das imagens associadas a identificação dos organismos. Foi determinado a melhor forma de trabalho com o Sino, necessidade de acessórios e adaptações e protocolo de uso, assim como a relação entre a transparência da água (disco de Secchi) e a possibilidade de identificação e quantificação de formas planctônicas nas imagens. Os resultados apontam para o uso do Sino em baixa velocidade, ou seja, em perfilagem vertical e com iluminação apresentando os melhores resultados e maior eficiência na identificação das ocorrências de organismos. A relação entre a transparência da água e a obtenção de imagens de boa qualidade foram observadas, de forma mais segura, em condições de Secchi superior a 2,5 m. Em função das condições de transparência da água e da variabilidade de ocorrência dos organismos a probabilidade de registro de organismos foi de 37,5% do total de saídas a campo. Por fim, as análises dos vídeos indicaram um possível comportamento de migração vertical de Olindias sambaquiensis (Hydrozoa) e chamaram a atenção pela abundância de Pseudo-TEPs (Transparent Exopolymer Particles), que acredita-se são de grande importância na ecologia pelágica de águas costeiras
Inteligência Artificial na Execução Fiscal transformação tecnológica e governança regulatória
Get PDFABSTRACTTax enforcement is one of the biggest challenges facing the BrazilianJudiciary due to the large number of processes and the complexityof credit recovery. This study explores how Artificial Intelligence(AI), by automating repetitive tasks and assisting in decision making,can increase the efficiency and speed of tax executions inthe context of Brazil, while at the same time presenting risks andlimitations, such as the possibility of algorithmic biases and lack oftransparency in automated decisions. The methodology adopted inthis work was an integrative literature review with the analysis ofscientific articles and reports that investigate the impact of ArtificialIntelligence techniques on tax enforcement. The results revealthat, although Artificial Intelligence has the potential to transformtax enforcement, increasing effectiveness in credit recovery andoptimizing procedural management, its implementation still facesobstacles, such as the need for regulation and technological integrationwith judicial systems. existing. It is concluded that, for AI to beapplied effectively and ethically, it is essential to develop a robustregulatory framework that ensures continuous human supervision,transparency of algorithms and the protection of rights. Thus, thestudy suggests that the combination of technological innovationwith adequate governance is essential to maximize the benefits ofAI and mitigate its restrictions in the context of tax executions
Relato de Experiência: Aprendizagem Ativa no Ensino de Visão Computacional com Cascade e HOG
Get PDFABSTRACTComputer vision, a growing field in artificial intelligence, is crucialin sectors such as healthcare and security. This article describes theapplication of active teaching methodology in a short course oncomputer vision, involving 26 Software Engineering students. Thecourse covered the techniques of Cascade Classifier and Histogramof Oriented Gradients (HOG), integrating theory and practice over20 hours of activities. The effectiveness of the course was evaluatedthrough quantitative and qualitative analyses. Quantitativeresults showed a significant improvement in students’ performance,while qualitative analysis indicated an increase in practical skillsand understanding of the techniques, despite challenges related totheoretical workload and activity support. The practical approachwas praised, with suggestions to improve the teaching materialsand support during activities
PROV-SwSystem: AWeb System to Capture and Analyze Provenance Data from Software Development Processes
Get PDFABSTRACTData provenance encompasses the description of the origin andhistorical events of data processing. Although models such as W3CPROV allow capturing and analyzing provenance data, additionalmethods are required to address software development processes’specific uses and complexities. In this context, the PROV-SwProcessprovenance model was developed to extend W3C PROV. However,the PROV-SwProcess provenance model does not provide a specifictool to enable its instantiation. In this vein, this paper aims topresent the PROV-SwSystem - a web-based system to capture, store,and analyze software process data along with their provenance,adhering to the PROV-SwProcess model. To achieve this, the DesignScience Research (DSR) methodology was adopted, supportingthe system creation, construction, and validation. PROV-SwSystemwas implemented using Javascript, Bootstrap, Node.js, and Jquery,encompassing all predefined use cases. Unit tests and system testswere performed, identifying six redundant variables and five redundantfunctions. In addition, an exploratory usability test wasperformed with an external user, revealing fifteen navigation issues,eight attempts to register data in inappropriate sections, sixregistration confirmation failures, and data integration problems.Actions for data entry validation and interface improvements wereimplemented, enabling the tool’s usage and paving the way forfuture testing on a larger scale
Análise de Desempenho de Frameworks para Aplicações Distribuídas
Get PDFAbstractAs distributed systems become increasingly pervasive in everydaylife, performance analyses that provide quantifiable insights intothe efficiency of these systems are essential. This study proposesan analysis of the performance of backend frameworks that enablethe development of such distributed systems. It is expected that thefindings of this project will guide better technology choices for thedevelopment of future distributed applications. The results indicatethat the choice of the ideal framework depends on the specific needsof the application. The Flask framework stands out as a robust andefficient option for scenarios requiring a high number of requestsand handling of lightweight data. Actix is an interesting choice forapplications that prioritize low resource consumption. ASP.NETcan be a good option for handling large volumes of data, despite itstendency for high CPU usage
Reconhecimento de Voz Aplicado ao Jogo de Xadrez Usando a Plataforma Arduino
Get PDFAbstractChess is a strategic game that enhances cognitive skills, yet itsphysical demands can exclude players with motor disabilities. Thisstudy presents a voice recognition system that enables chess piecemovement through verbal commands, enhancing accessibility. Theproposed solution integrates the VoiceRecognitionV3 module withthe Arduino Uno to recognize spoken commands related to gameactions. The methodology involves training the system to interpreta structured set of voice commands, organizing the chessboard intoquadrants for efficient processing. Initial tests confirm the feasibilityof voice-controlled chess interactions, demonstrating potentialbenefits for accessibility. However, limitations include the module’sconstraint of recognizing only seven commands simultaneously andthe necessity for user-specific voice training. Future improvementsmay involve integrating advanced speech recognition algorithmsand machine learning to expand command recognition and adaptability.This research highlights the potential of voice recognition ininteractive and inclusive gaming environments
Robótica Artística e Sustentável: Fortalecendo a Participação Feminina na Tecnologia
Get PDFABSTRACTEste trabalho relata a participação de três alunas do curso técnico integrado em Informática ao Ensino Médio, em um projeto de Robótica Aplicada. As alunas competiram na fase estadual da Olimpíada Brasileira de Robótica em 2024, na modalidade artística, e avançaram para a etapa nacional. O principal objetivo do projeto foi desenvolver robôs sustentáveis por meio da reutilização criativa de materiais recicláveis e componentes eletrônicos, visando reduzir os impactos ambientais associados ao descarte inadequado desses itens. A iniciativa adotou a metodologia de aprendizado baseado em projetos, que promoveu o desenvolvimento de habilidades práticas em eletrônica e programação. Além de fomentar a sustentabilidade, o projeto também incentivou a participação feminina nas inovações tecnológicas. Como resultado, as alunas não apenas aprimoraram suas habilidades técnicas e criativas, mas também conquistaram o primeiro lugar na etapa estadual da OBR. Além disso, desempenharam um papel crucial na promoção da conscientização ambiental e na ampliação da visibilidade feminina no campo da tecnologi
Um Sistema para Monitoramento Remoto de Sinais de Eletrocardiograma em Ambientes Hospitalares e de Home Care
Get PDFABSTRACTAnnually, about 17.9 million people in the world die from someform of cardiovascular disease, and most of these deaths could beprevented with a simple ECG (electrocardiogram) exam, followed byan adequate medical treatment. The ECG is an exam that collects theheart’s electrical impulses. These data can be visualized in graphsthat allow the identification of possible alterations or behaviors outof the known parameters, that may cause harm to the patients. Theproposed systems enables medical team to collect ECG signals anddetect some anomaly in patients allocated in hospitals and homecare environments. The data collected are transmitted through aWi-Fi network to a server at the internet cloud, where they arepropperly storaged. This way, the medical team can visualize thenecessary parameters in real-time using an App designed to mobiledevices. Real experiments were performed, in a case study, showingthe feasibility of the proposed system
Fogo Virtual, Impacto Real: Educação Ambiental e Preservação Amazônica com Realidade Aumentada
Get PDFABSTRACTThe fires in the Amazon represent an environmental crisis that threatens biodiversity and global climate balance, affecting local communities and contributing to climate change. Environmental awareness, especially among young people, is crucial to tackling this issue. Therefore, an innovative project is being developed to raise awareness among elementary school students about the fires in a region. The initiative involves creating an educational book with information on forest fires, using augmented reality (AR) to simulate fires in 3D and display videos in the cultural settings of a specific locality. This approach aims to educate and sensitize children, promoting the preservation of the forest and encouraging environmental responsibility from an early age
JOGO EM REALIDADE VIRTUAL DE TREINAMENTO PARA COMPENSAÇÃO DE PERDA DO CAMPO VISUAL PARA PACIENTES PÓS-AVC
Get PDFAbstractStroke, or Cerebrovascular Accident, affects thousands of individualsannually and remains a leading cause of death in Brazil. Amongsurvivors, many experience lasting sequelae, such as hemianopsia—a partial loss of the visual field. The standard treatment for thiscondition typically involves visual field compensation exercises,which, while effective, are often repetitive and demotivating forpatients. To address these challenges, we developed AeroVR, aninnovative game designed to support patient rehabilitation. AeroVRcombines virtual reality technology, which facilitates the stimulationof a broader visual field, with the principles of serious gamesto enhance patient engagement and adherence to therapy. Developediteratively in partnership with the Occupational TherapySchool-Clinic, the game is highly customizable, ensuring it cancater to a diverse range of patients. By merging advanced technologywith user-centered design, AeroVR offers a promising solutionfor improving the effectiveness and accessibility of hemianopiarehabilitation