DISSERTAÇÃO DE MESTRADO - PROGRAMA DE PÓS-GRADUAÇÃO EM CIÊNCIA DA COMPUTAÇÃO
URI Permanente para esta coleçãohttps://tedebc-teste.ufma.br/handle/tede/1314
Áreas de Concentração e Linhas de Pesquisa:
Automação e Contrôle
Ciência da Computação
Sistemas de Energia Elétrica
Navegar
Navegando DISSERTAÇÃO DE MESTRADO - PROGRAMA DE PÓS-GRADUAÇÃO EM CIÊNCIA DA COMPUTAÇÃO por Autor "ALMEIDA, João Dallyson Sousa de"
Agora exibindo 1 - 3 de 3
- Resultados por página
- Opções de Ordenação
Item Aprendizagem Profunda Aplicada ao Diagnóstico de Melanoma(Universidade Federal do Maranhão, 2019-02-14) MAIA, Lucas Bezerra; PAIVA, Anselmo Cardoso de; 375523843-87; http://lattes.cnpq.br/6446831084215512; BRAZ JÚNIOR, Geraldo; 000520303-18; http://lattes.cnpq.br/8287861610873629; BRAZ JÚNIOR, Geraldo; 000520303-18; http://lattes.cnpq.br/8287861610873629; PAIVA, Anselmo Cardoso de; 375523843-87; http://lattes.cnpq.br/6446831084215512; ALMEIDA, João Dallyson Sousa de; http://lattes.cnpq.br/6047330108382641; CARVALHO FILHO, Antonio Oseas de; http://lattes.cnpq.br/7913655222849728Melanoma is the most lethal type of cancer when compared to others skin diseases. However, when the diagnosis is made in its initial stage, patients have high rates of recovery. Several approaches to automatic detection and diagnosis of melanoma have been explored by different authors in order to provide an auxiliary opinion to specialists. Training models with the existing data sets have been a difficult task due to the problem of imbalanced data. This work aims to evaluate to the evaluation the performance of machine learning algorithms combined with imbalanced learning technique, regarding the task of melanoma diagnosis. The architectures of Convolutional Neural Networks VGG16, VGG19, Inception, and ResNet were used along with ABCD rule to extract patterns of skin lesions in a set of 200 dermatoscopic images. The Random Forest classifier reached a sensitivity of 92.5 % and a kappa index of 77.15 % after the use of attribute selection with Greedy Stepwise and balancing the training data with Synthetically Minority Oversampling TEchnique (SMOTE) and the Edited Nearest Neighbor (ENN) rule.Item Aprendizagem Profunda Aplicada ao Diagnóstico do Glaucoma(Universidade Federal do Maranhão, 2019-02-19) LIMA, Alan Carlos de Moura; ALMEIDA, João Dallyson Sousa de; 003998573-38; http://lattes.cnpq.br/6047330108382641; BRAZ JÚNIOR, Geraldo; 000520303-18; http://lattes.cnpq.br/8287861610873629; BRAZ JÚNIOR, Geraldo; 000520303-18; http://lattes.cnpq.br/8287861610873629; ALMEIDA, João Dallyson Sousa de; http://lattes.cnpq.br/6047330108382641; PAIVA, Anselmo Cardoso de; http://lattes.cnpq.br/6446831084215512; VERAS, Rodrigo de Melo Souza; http://lattes.cnpq.br/2634254790193199Glaucoma is a cluster of ocular diseases that cause damage to the eye’s optic nerve and cause successive narrowing of the visual field in affected patients, due to an increase in intraocular pressure, which can lead the patient to blindness at an advanced stage without clinical reversal. For several years, from techniques of manual analysis of the internal structures of the eye to the use of deep learning with convolutional neural networks (CNNs) were successfully used in the diagnosis of glaucoma. However, building a deep learning network requires a lot of effort that in many situations is not always able to achieve satisfactory results due to the amount of parameters that need to be configured to adapt the CNN architecture to the problem in question. The objective of this work is to use a hyperparameter search technic to select the tuned parameters of a genetic algorithm (GA) to select the best CNN architecture through evolutionary techniques and to be able to aid in the accurate diagnosis of glaucoma, in eye fund images. The proposed methodology was applied in 455 images from RIM-ONE dataset, in its version 2 (r2), with resized images to 96x96 pixels in the RGB color model. The selected CNN by AG, after its training, achieved for the diagnosis of glaucoma the results of 96.63% for accuracy, 94.87% for sensitivity, 98.00% for specificity, 97.37% for precision and 96.10% for f-score.Item Redes neurais multinível para classificação do ângulo da câmara anterior utilizando Imagens OCT-SA(Universidade Federal do Maranhão, 2021-03-18) FERREIRA, Marcos Melo; BRAZ JUNIOR, Geraldo; http://lattes.cnpq.br/8287861610873629; BRAZ JUNIOR, Geraldo; http://lattes.cnpq.br/8287861610873629; PAIVA, Anselmo Cardoso de; http://lattes.cnpq.br/6446831084215512; ALMEIDA, João Dallyson Sousa de; http://lattes.cnpq.br/6047330108382641; ARAÚJO, Sidnei Alves de; http://lattes.cnpq.br/2542529753132844Glaucoma is identified as one of the main causes of visual impairment, and the main cause of irreversible blindness. The main forms of the disease are primary open-angle glaucoma and primary angle-closure glaucoma. In people with angle-closure glaucoma, the anterior chamber angle narrows, consequently causing an increase in intraocular pressure causing damage to the optic nerve, causing partial or total vision loss. As the damage is irreversible, an early diagnosis is essential, but it is hampered due to the fact that the disease is asymptomatic in early stages. For early detection of the disease, routine imaging tests are recommended, one of which is Anterior Segment Optical Coherence Tomography, which allows an angle classification, which is essential for diagnosis. An analysis of this type of image requires a degree of interpretation on the part of specialists, because of this, the evaluation of many images requires a lot of time, which can lead to professional fatigue. The use of automated methods to assist in the interpretation of images would contribute to get diagnoses more quickly. In this work, an automated method is proposed to classify the anterior chamber angle, present in Anterior Segment images, based on deep learning, using convolutional neural networks. Initially, five pre-trained models of convolutional networks were adjusted to perform feature extraction and classify images. Next, the models were combined in a multilevel architecture, with the objective of increasing the classification capacity. As best result achieved an AUC value (Area Under the Curve) of 0.999.