Mestrado em Ciência da Computação (INF)

URI Permanente para esta coleção

http://200.137.215.59/tede/handle/tde/225

Navegar

Agora exibindo 1 - 9 de 9

Classificação de documentos da administração pública utilizando inteligência artificial
(Universidade Federal de Goiás, 2024-04-30) Carvalho, Rogerio Rodrigues; Costa, Ronaldo Martins da; http://lattes.cnpq.br/7080590204832262; Costa, Ronaldo Martins da; Souza, Rodrigo Gonçalves de; Silva, Nádia Félix Felipe da
Public organizations face difficulties in classifying and promoting transparency of the numerous documents produced during the execution of their activities. Correct classification of documents is critical to prevent public access to sensitive information and protect individuals and organizations from malicious use. This work proposes two approachs to perform the task of classifying sensitive documents, using state-of-the-art artificial intelligence techniques and best practices found in the literature: a conventional method, which uses artificial intelligence techniques and regular expressions to analyze the textual content of documents, and an alternative method, which employs the CBIR technique to classify documents when text extraction is not viable. Using real data from the Electronic Information System (SEI) of the Federal University of Goiás (UFG), the results achieved demonstrated that the application of regular expressions as a preliminary check can improve the computational efficiency of the classification process, despite showing a modest increase in classification precision. The conventional method proved to be effective in document classification, with the BERT model standing out for its performance with an accuracy rate of 94%. The alternative method, in turn, offered a viable solution for challenging scenarios, showing promising results with an accuracy rate of 87% in classifying public documents
Construção de dispositivo com software embarcado de avaliação do reflexo pupilar humano para apoio a diagnóstico oftalmológico
(Universidade Federal de Goiás, 2021-08-06) Delfino, Higor Pereira; Costa, Ronaldo Martins da; http://lattes.cnpq.br/7080590204832262; Costa, Ronaldo Martins da; Gonçalves, Cristhiane; Laureano, Gustavo Teodoro
The realization of diagnosis through precise metrics, in a fast, safe, efficient way, with non-evasive means and at a low cost, brings great advantages for the medical field. Meeting this need are studies in the field of the human pupillary reflex that demonstrate that the eyes are more than an organ of the sensory system and can provide accurate and reliable biosignals to aid in diagnosis. This work aims to build an automated pupillometry equipment in order to help identify pathologies or disorders, using computer vision techniques. In this sense, this work proposed a pupilometer that can be used in people, capable of stimulating the pupil at various wavelengths, providing a friendly interface and pre-assessing the exam. This work proposes to build a low-cost, easy-to-operate and minimally viable equipment to be used in an ophthalmological office. The equipment built in this work presents adequate quality, automation and configuration flexibility with a great potential of use for several studies in the field of automated pupillometry. Regarding the induction of pupillary reactivity, the equipment is capable of working with flexible and dynamic configurations, and it can be adjusted to the intensity of stimuli and worked with strobe light.
Um sistema WebGIS para classificação supervisionada de cobertura do solo utilizando inteligência artificial
(Universidade Federal de Goiás, 2022-10-21) Fernandes, Yuri Kuivjogi; Costa, Ronaldo Martins da; http://lattes.cnpq.br/7080590204832262; Costa, Ronaldo Martins da; Oliveira, Bruna Mendes de; Cremon, Édipo Henrique
With the advancement in data generation for Earth observation and its availability free of charge, the Remote Sensing (SR) area advanced significantly. Over the years, it has been observed the migration of RS applications to the internet environment, facilitating searches of different uses. This work proposes a new approach for collecting and manipulating spatial data for spectral classification based on pixels. A web application was built integrating Google Earth Engine, Google Maps and Auto Machine Learning services for performance analysis. Experiments using samples from land cover regions in Goiás, Brazil, justifying the gain in time, processing and data storage. Such contributions are related to the large amount of information from satellite images collected in a conventional way, which are later not used. As a final result, there is an image classified through the classification process representing the different land cover classes. Model training achieved an accuracy of 99.85% using the Light Gradient Boosting Machine (LightGBM) model. In addition to these benefits, the optimization of processes allows the inclusion of research from other major areas, thus for the greater dissemination of knowledge in the area of SR and pattern recognition applications.
Identificação e estimativa da altura de árvores em imagens de satélite e do Google Street View
(Universidade Federal de Goiás, 2016-12-20) Lima, Heuber Gustavo Frazao de; Costa, Ronaldo Martins da; http://lattes.cnpq.br/7080590204832262; Soares, Fabrízzio Alphonsus Alves de Melo Nunes; Pires, Sandrerley Ramos
The electrical distribution is a critical activity since many people depend of this service. Faults in the distribution system occur from several factors that can damage the system and therefore interrupt the supply of energy. Among the various factors that may cause problems this work proposes a automatic detection of trees near or even in the distribution network. In order to avoid that the trees to force or even rupture of the distribution cables, are made the pruning of the trees that have some kind of risk to the network. However, this activity is usually manual and teams must sift through all the network for problems. The main objective of this work is to propose a process, based on computer vision, which allows the automated identification of nearby trees or under the power distribution network from aerial images provided by Google Earth and even estimate the height of the same from 2D Google Street View images.
Aplicação de técnicas de content-based image retrieval (CBIR) em imagens radiográficas
(Universidade Federal de Goiás, 2016-09-30) Macena Júnior, Elias Borges; Costa, Ronaldo Martins da; http://lattes.cnpq.br/7080590204832262; Costa, Ronaldo Martins da; Oliveira, Leandro Luís Galdino de; Silva, Fernanda Paula Yamamoto
In order to improve the diagnostic process several research centers have focused on the development of information systems applying powerful techniques of computer-aided diagnosis (CAD). In this context, the creation of content-based image retrieval (CBIR) is an important step in developing an efficient CAD system. This work proposes the validation of recovery with a hybrid CBIR method based on 2D medical images. The results of the techniques applied, indicate a hit rate of 90.25% and indicate a gain of 35% in the performance of techniques, that is, the time search and retrieval of images, paving the way for the development of information systems more efficient to build support generic diagnostic systems.
Detecção da alcoolemia por meio da pupilometria dinâmica
(Universidade Federal de Goiás, 2016-07-01) Pinheiro, Hedenir Monteiro; Costa, Ronaldo Martins da; http://lattes.cnpq.br/7080590204832262; Costa, Ronaldo Martins da; Laureano, Gustavo Teodoro; Melo, Cinthia Mendonça de
The consumption of alcohol beverages causes disturbance to society, such as traffic accidents, problems at work and interpersonal violence. The main methods for the detection of alcohol are the use of a breathalyzer and a blood test. Such methods are, however, relatively invasive. The purpose of this work is to develop a portable device and a method capable of performing the dynamic pupillometry to verify the feasibility of detect alcohol use by means of recording pupillary movements under light stimulation. After the device creation, was recorded and collected biometric information of 49 people in sober and inebriated state, producing a database with 356 video. So were selected the most relevant pupillary features. This set of features was subjected to statistical tests and pattern recognition algorithms KNN, Ensemble KNN and SVM. The results showed that pupillometry was performed efficiently with 96% accuracy, the alcohol changes the pupillary behavior and detecting alcohol consumption is feasible when the analysis is done individually and with many videos.
Reconhecimento do tipo de cachaça utilizando visão computacional e reconhecimento de padrões
(Universidade Federal de Goiás, 2015-10-01) Rodrigues, Bruno Urbano; Soares, Anderson da Silva; http://lattes.cnpq.br/1096941114079527; Costa, Ronaldo Martins da; http://lattes.cnpq.br/7080590204832262; Costa, Ronaldo Martins da; Silva, Anderson Soares; Salvini, Rogério Lopes; Caliari, Marcio
The cachaça is a type of drink distilled from sugar cane that has a great economic importance. Their classification includes three types: aged, premium and premium extra. These three classifications are related to the aging time drink in wooden barrels. Besides the aging time is relevant to know what the wood used in the barrels of storage for the properties of each drink are informed correctly to the consumer. This dissertation presented a method for the automatic recognition of the type of wood and the aging time using a computer vision system. The computer vision system is used in the analysis of the color models (RGB) additive and subtractive (CIELab) caught on digital camera. In association with computer vision, algorithmics, system of pattern recognition are used in conjunction with chemical information for the classification of samples. Went used four algorithmics: Artificial Neural network, k-NN (k-Nearest Neighbor), SVM (Support Vector Machines) and Naive Bayes. The end is used the ensemble AdaBoost, technique combining classifiers. In the study we used 108 samples of rum. The results obtained show that it was possible to obtain rates excess use of % 96.26 algorithmics of pattern recognition to the problem of the type of wood. The AdaBoost brought 100 indices % hit to the problem of classification of the type of wood and aging time. Your use proves that it is possible the sort of rum using only color model data contributing to a lower cost of production.
Metodologia para análise de imagens de baixa resolução, para definição de MUB (Mapa Urbano Básico) para apoio às concessionárias de distribuição
(Universidade Federal de Goiás, 2018-05-03) Santos, Paulo Victor dos; Gonçalves, Cristhiane; http://lattes.cnpq.br/3935775322457150; Costa, Ronaldo Martins da; http://lattes.cnpq.br/7080590204832262; Costa, Ronaldo Martins da; Gonçalves, Cristhiane; Soares, Fabrízzio Alphonsus Alves de Melo Nunes; Matsushima, Luciana Cardoso
The high cost of geo-referenced technologies capable of keeping up-to-date geographic information system data from a power distribution company, which has the basic urban map as the primary data layer, may result in non-purchase of this product, making information outdated or incomplete, generating in addition to losses, rework and confusion when there is a need to do verifications and validations that could be performed remotely. This proposal to support energy distribution concessionaires will use Computational Vision (VC) and Digital Image Processing (PDI) methods, allowing a low cost and efficient maintenance in layers of buildings present in basic urban maps of distributors. This findings might result in possible savings, without any need of displacement in the field to observe a situation that can be evidenced remotely.
Pupilometria na investigação de diabetes mellitus tipo II
(Universidade Federal de Goiás, 2018-09-28) Silva, Cleyton Rafael Gomes; Gonçalves, Cristhiane; http://lattes.cnpq.br/3935775322457150; Costa, Ronaldo Martins da; http://lattes.cnpq.br/7080590204832262; Costa, Ronaldo Martins da; Gonçalves, Cristhiane; Salvini, Rogério Lopes; Taleb, Alexandre Chater
Examining human pupillary behavior is a non-invasive, low-cost method for assessing neurological activity. Changes in this behavior are correlated to various health conditions, such as: Parkinson’s, Alzheimer’s, autism and diabetes. In order to obtain information about the pupillary behavior, it is necessary to measure the pupil diameter in procedures that induce pupillary reflexes, known as Pupilometry. Pupillary measurement is made by filming the procedures when applying computer vision techniques for pupil recognition. The objective of this research was to develop an Automated Pupilometry System (SAP) to support the investigation of patients with type II diabetes mellitus. SAP was able to record, induce, and extract 96 pupil features. In the experiment with 15 healthy patients and 16 diabetics, a 94% accuracy in the identification of diabetics type II was obtained, demonstrating the efficiency of SAP for the performance of examinations, and evidencing the potential of pupil use in the investigation of diabetes mellitus type II.

Navegar

Navegando Mestrado em Ciência da Computação (INF) por Por Orientador "Costa, Ronaldo Martins da"

Resultados por página

Opções de Ordenação