Eliana Almeida | UFAL - Universidade Federal de Alagoas (original) (raw)
Papers by Eliana Almeida
Anais Estendidos do XXXV Conference on Graphics, Patterns and Images (SIBGRAPI Estendido 2022), Oct 24, 2022
Anais do XV Women in Information Technology (WIT 2021), 2021
O presente artigo apresenta relatos a respeito do projeto de extensão da Universidade Federal de ... more O presente artigo apresenta relatos a respeito do projeto de extensão da Universidade Federal de Alagoas intitulado: "Katie: saindo do buraco negro e impulsionando as meninas para a computação", que visa motivar, apoiar e promover a inclusão das mulheres nas áreas de Ciência, Tecnologia, e Engenharia e Matemática (STEM). O artigo descreve como o projeto Katie permaneceu ativo durante o período de pandemia ocasionado pelo COVID-19. Para que as atividades ocorressem, foram utilizadas ferramentas online, como as mídias sociais e as plataformas digitais para levar ao público uma inclusão de assuntos relacionados à Tecnologia da Informação, além de ser tema de um artigo publicado em um evento nacional da área de comunicação.
Anais Principais do Simpósio Brasileiro de Computação Aplicada à Saúde (SBCAS 2020), 2020
Doenças Pulmonares Intersticiais (DPIs) são caracterizadas pela cicatrização progressiva do inter... more Doenças Pulmonares Intersticiais (DPIs) são caracterizadas pela cicatrização progressiva do interstício pulmonar e podem levar a insuficiência respiratória. Este artigo propõe um método de classificação de DPIs a partir de imagens de Tomografia Computadorizada (TC) mapeadas em uma Rede Complexa. Métricas de centralidade foram usadas com o objetivo de obter seus atributos texturais. Utilizando um classificador KNN, os resultados apresentaram uma acurácia média de 89.81%. Para os padrões de textura de DPI do tipo consolidação pulmonar e opacidade em vidro fosco, a acurácia do método foi de 90% e 86%, respectivamente, o que aponta o método proposto como promissor para estudos futuros em imagens de TC associadas a pacientes com COVID-19.
Anais do Simpósio Brasileiro de Computação Aplicada à Saúde (SBCAS), 2020
O objetivo deste trabalho é auxiliar no desenvolvimento de uma ferramenta de diagnóstico de doenç... more O objetivo deste trabalho é auxiliar no desenvolvimento de uma ferramenta de diagnóstico de doenças pulmonares auxiliado por computador. Nessa primeira etapa utilizamos análise de componentes principais (PCA), análise do discriminante linear (LDA) e o algoritmo de k-vizinhos mais próximos (KNN) para classificar 3252 regiões de interesse (ROI) de Tomografias Computadorizadas de Alta-Resolução de tórax em relação à 6 padrões pulmonares. Cada ROI possui um total de 28 dimensões que foram reduzidas por PCA e LDA e então classificadas por KNN (k = 5). Obtivemos uma taxa de classificação correta de 80,82% em 13 dimensões com PCA e 83,74% em 5 dimensões com LDA.
Methods of information in medicine, Jan 8, 2018
Diffuse lung diseases (DLDs) are a diverse group of pulmonary disorders, characterized by inflam... more Diffuse lung diseases (DLDs) are a diverse group of pulmonary disorders, characterized by inflammation of lung tissue, which may lead to permanent loss of the ability to breathe and death. Distinguishing among these diseases is challenging to physicians due their wide variety and unknown causes. Computer-aided diagnosis (CAD) is a useful approach to improve diagnostic accuracy, by combining information provided by experts with Machine Learning (ML) methods. Exploring the potential of dimensionality reduction combined with ML methods for diagnosis of DLDs; improving the classification accuracy over state-of-the-art methods. A data set composed of 3252 regions of interest (ROIs) was used, from which 28 features were extracted per ROI. We used Principal Component Analysis, Linear Discriminant Analysis, and Stepwise Selection - Forward, Backward, and Forward-Backward to reduce feature dimensionality. The feature subsets obtained were used as input to the following ML methods: Support...
2015 37th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), 2015
We propose the use of fuzzy membership functions to analyze images of diffuse pulmonary diseases ... more We propose the use of fuzzy membership functions to analyze images of diffuse pulmonary diseases (DPDs) based on fractal and texture features. The features were extracted from preprocessed regions of interest (ROIs) selected from high-resolution computed tomography images. The ROIs represent five different patterns of DPDs and normal lung tissue. A Gaussian mixture model (GMM) was constructed for each feature, with six Gaussians modeling the six patterns. Feature selection was performed and the GMMs of the five significant features were used. From the GMMs, fuzzy membership functions were obtained by a probability-possibility transformation and further statistical analysis was performed. An average classification accuracy of 63.5% was obtained for the six classes. For four of the six classes, the classification accuracy was superior to 65%, and the best classification accuracy was 75.5% for one class. The use of fuzzy membership functions to assist in pattern classification is an alternative to deterministic approaches to explore strategies for medical diagnosis.
2015 IEEE International Symposium on Medical Measurements and Applications (MeMeA) Proceedings, 2015
This paper presents results of statistical analysis of fractal and texture features obtained from... more This paper presents results of statistical analysis of fractal and texture features obtained from images of diffuse pulmonary diseases (DPDs). The features were extracted from preprocessed regions of interest (ROIs) selected from high-resolution computed tomography images. The ROIs represent five different patterns of DPDs and normal lung tissues. A Gaussian mixture model (GMM) was constructed for each feature, including all patterns. For each GMM, the six classes were identified and compared with the radiological classification of the corresponding ROIs. In 78.5% of the features, the GMM provides, for at least one class, a correct classification of at least 60%. The GMM approach facilitates detailed statistical analysis of the characteristics of each feature and assists in the development of classification strategies.
Proceedings. 17th Brazilian Symposium on Computer Graphics and Image Processing
Abstract This paper presents the image formation process in vibro-acoustography for systems based... more Abstract This paper presents the image formation process in vibro-acoustography for systems based on sector array transducers. These transducers are an alternative to annular concave transducers. They represent an innovative technique deserving detailed ...
Journal of Network and Computer Applications, 2014
ABSTRACT This work proposes an embedded system that opportunistically detects free on-street park... more ABSTRACT This work proposes an embedded system that opportunistically detects free on-street parking by using surveillance cameras. Intelligent boards are embedded in the cameras allowing a distributed processing and communication. The task of looking for free parking spaces demands a significant amount of drivers' time. This problem generally takes place in urban areas where there is a large number of vehicles and few available parking spaces (on-street or off-street). Decreasing the time spent in this task could reduce the traffic jam or, in an extreme perspective, alleviate the toxic gases emission or the fossil combustible consumption. Initially, it was evaluated three classical techniques for image processing applied to detect free on-street parking. The best one presents a success detection rate close to 100%. After that, it was evaluated the performance of the embedded system by using only the best image processing technique. This performance evaluation consider three different scenarios: centralized, hybrid, and embedded. The last one is the main proposal and contribution. The results reveal that embedded proposal had an average requisition time of 2.10 s vs. 0.38 s of centralized one. The hybrid one presents the worst results.
This paper presents and discusses the use of a new feature for PolSAR imagery: the Generalized St... more This paper presents and discusses the use of a new feature for PolSAR imagery: the Generalized Statistical Complexity. This measure is able to capture the disorder of the data by means of the entropy, as well as its departure from a reference distribution. The latter component is obtained by measuring a stochastic distance between two models: the mathcalG0\mathcal G^0mathcalG0 and the Gamma laws. Preliminary results on the intensity components of AIRSAR image of San Francisco are encouraging.
The Visual Computer, 2007
Computational & Applied Mathematics, 2012
Brazilian Journal of Probability and Statistics, 2009
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 2012
In 2018, only 10% of students enrolled in Computer Science and Computer Engineering courses at th... more In 2018, only 10% of students enrolled in Computer Science and Computer Engineering courses at the Federal University of Alagoas (UFAL) were women, one of the issues that highlight female underrepresentation in the academic space. In this scenario, the Katie Group appears, formed by the students of these courses, focused on initiating the process of reversing this low female representation, acting not only in academic spaces, but also in these previous educational environments, such as high school. So, with that inspiring goal in mind, the group was named after Katherine Bouman, the computer scientist responsible for the algorithm used to create the first image of a huge black
Brazilian Journal of Probability and Statistics, 2009
The Visual Computer, 2007
In this work we propose a new approach for fast visualization and exploration of virtual worlds b... more In this work we propose a new approach for fast visualization and exploration of virtual worlds based on the use of cartographic concepts and techniques. Versions of cartographic maps with different levels of details can be created by using a set of operations named cartographic generalization. Cartographic generalization employs twelve operators and domain-specific knowledge, being the contribution of this work their transposition to 3D virtual worlds. The architecture of a system for 3D generalization is proposed and the system is implemented. Differently from traditional cartographic processes, we use artificial intelligence for both selecting the key objects and applying the operators. As a case study, we present the simplification of the historical quarter of Recife (Brazil).
Abstract This paper presents the image formation process in vibro-acoustography for systems based... more Abstract This paper presents the image formation process in vibro-acoustography for systems based on sector array transducers. These transducers are an alternative to annular concave transducers. They represent an innovative technique deserving detailed ...
Anais Estendidos do XXXV Conference on Graphics, Patterns and Images (SIBGRAPI Estendido 2022), Oct 24, 2022
Anais do XV Women in Information Technology (WIT 2021), 2021
O presente artigo apresenta relatos a respeito do projeto de extensão da Universidade Federal de ... more O presente artigo apresenta relatos a respeito do projeto de extensão da Universidade Federal de Alagoas intitulado: "Katie: saindo do buraco negro e impulsionando as meninas para a computação", que visa motivar, apoiar e promover a inclusão das mulheres nas áreas de Ciência, Tecnologia, e Engenharia e Matemática (STEM). O artigo descreve como o projeto Katie permaneceu ativo durante o período de pandemia ocasionado pelo COVID-19. Para que as atividades ocorressem, foram utilizadas ferramentas online, como as mídias sociais e as plataformas digitais para levar ao público uma inclusão de assuntos relacionados à Tecnologia da Informação, além de ser tema de um artigo publicado em um evento nacional da área de comunicação.
Anais Principais do Simpósio Brasileiro de Computação Aplicada à Saúde (SBCAS 2020), 2020
Doenças Pulmonares Intersticiais (DPIs) são caracterizadas pela cicatrização progressiva do inter... more Doenças Pulmonares Intersticiais (DPIs) são caracterizadas pela cicatrização progressiva do interstício pulmonar e podem levar a insuficiência respiratória. Este artigo propõe um método de classificação de DPIs a partir de imagens de Tomografia Computadorizada (TC) mapeadas em uma Rede Complexa. Métricas de centralidade foram usadas com o objetivo de obter seus atributos texturais. Utilizando um classificador KNN, os resultados apresentaram uma acurácia média de 89.81%. Para os padrões de textura de DPI do tipo consolidação pulmonar e opacidade em vidro fosco, a acurácia do método foi de 90% e 86%, respectivamente, o que aponta o método proposto como promissor para estudos futuros em imagens de TC associadas a pacientes com COVID-19.
Anais do Simpósio Brasileiro de Computação Aplicada à Saúde (SBCAS), 2020
O objetivo deste trabalho é auxiliar no desenvolvimento de uma ferramenta de diagnóstico de doenç... more O objetivo deste trabalho é auxiliar no desenvolvimento de uma ferramenta de diagnóstico de doenças pulmonares auxiliado por computador. Nessa primeira etapa utilizamos análise de componentes principais (PCA), análise do discriminante linear (LDA) e o algoritmo de k-vizinhos mais próximos (KNN) para classificar 3252 regiões de interesse (ROI) de Tomografias Computadorizadas de Alta-Resolução de tórax em relação à 6 padrões pulmonares. Cada ROI possui um total de 28 dimensões que foram reduzidas por PCA e LDA e então classificadas por KNN (k = 5). Obtivemos uma taxa de classificação correta de 80,82% em 13 dimensões com PCA e 83,74% em 5 dimensões com LDA.
Methods of information in medicine, Jan 8, 2018
Diffuse lung diseases (DLDs) are a diverse group of pulmonary disorders, characterized by inflam... more Diffuse lung diseases (DLDs) are a diverse group of pulmonary disorders, characterized by inflammation of lung tissue, which may lead to permanent loss of the ability to breathe and death. Distinguishing among these diseases is challenging to physicians due their wide variety and unknown causes. Computer-aided diagnosis (CAD) is a useful approach to improve diagnostic accuracy, by combining information provided by experts with Machine Learning (ML) methods. Exploring the potential of dimensionality reduction combined with ML methods for diagnosis of DLDs; improving the classification accuracy over state-of-the-art methods. A data set composed of 3252 regions of interest (ROIs) was used, from which 28 features were extracted per ROI. We used Principal Component Analysis, Linear Discriminant Analysis, and Stepwise Selection - Forward, Backward, and Forward-Backward to reduce feature dimensionality. The feature subsets obtained were used as input to the following ML methods: Support...
2015 37th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), 2015
We propose the use of fuzzy membership functions to analyze images of diffuse pulmonary diseases ... more We propose the use of fuzzy membership functions to analyze images of diffuse pulmonary diseases (DPDs) based on fractal and texture features. The features were extracted from preprocessed regions of interest (ROIs) selected from high-resolution computed tomography images. The ROIs represent five different patterns of DPDs and normal lung tissue. A Gaussian mixture model (GMM) was constructed for each feature, with six Gaussians modeling the six patterns. Feature selection was performed and the GMMs of the five significant features were used. From the GMMs, fuzzy membership functions were obtained by a probability-possibility transformation and further statistical analysis was performed. An average classification accuracy of 63.5% was obtained for the six classes. For four of the six classes, the classification accuracy was superior to 65%, and the best classification accuracy was 75.5% for one class. The use of fuzzy membership functions to assist in pattern classification is an alternative to deterministic approaches to explore strategies for medical diagnosis.
2015 IEEE International Symposium on Medical Measurements and Applications (MeMeA) Proceedings, 2015
This paper presents results of statistical analysis of fractal and texture features obtained from... more This paper presents results of statistical analysis of fractal and texture features obtained from images of diffuse pulmonary diseases (DPDs). The features were extracted from preprocessed regions of interest (ROIs) selected from high-resolution computed tomography images. The ROIs represent five different patterns of DPDs and normal lung tissues. A Gaussian mixture model (GMM) was constructed for each feature, including all patterns. For each GMM, the six classes were identified and compared with the radiological classification of the corresponding ROIs. In 78.5% of the features, the GMM provides, for at least one class, a correct classification of at least 60%. The GMM approach facilitates detailed statistical analysis of the characteristics of each feature and assists in the development of classification strategies.
Proceedings. 17th Brazilian Symposium on Computer Graphics and Image Processing
Abstract This paper presents the image formation process in vibro-acoustography for systems based... more Abstract This paper presents the image formation process in vibro-acoustography for systems based on sector array transducers. These transducers are an alternative to annular concave transducers. They represent an innovative technique deserving detailed ...
Journal of Network and Computer Applications, 2014
ABSTRACT This work proposes an embedded system that opportunistically detects free on-street park... more ABSTRACT This work proposes an embedded system that opportunistically detects free on-street parking by using surveillance cameras. Intelligent boards are embedded in the cameras allowing a distributed processing and communication. The task of looking for free parking spaces demands a significant amount of drivers' time. This problem generally takes place in urban areas where there is a large number of vehicles and few available parking spaces (on-street or off-street). Decreasing the time spent in this task could reduce the traffic jam or, in an extreme perspective, alleviate the toxic gases emission or the fossil combustible consumption. Initially, it was evaluated three classical techniques for image processing applied to detect free on-street parking. The best one presents a success detection rate close to 100%. After that, it was evaluated the performance of the embedded system by using only the best image processing technique. This performance evaluation consider three different scenarios: centralized, hybrid, and embedded. The last one is the main proposal and contribution. The results reveal that embedded proposal had an average requisition time of 2.10 s vs. 0.38 s of centralized one. The hybrid one presents the worst results.
This paper presents and discusses the use of a new feature for PolSAR imagery: the Generalized St... more This paper presents and discusses the use of a new feature for PolSAR imagery: the Generalized Statistical Complexity. This measure is able to capture the disorder of the data by means of the entropy, as well as its departure from a reference distribution. The latter component is obtained by measuring a stochastic distance between two models: the mathcalG0\mathcal G^0mathcalG0 and the Gamma laws. Preliminary results on the intensity components of AIRSAR image of San Francisco are encouraging.
The Visual Computer, 2007
Computational & Applied Mathematics, 2012
Brazilian Journal of Probability and Statistics, 2009
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 2012
In 2018, only 10% of students enrolled in Computer Science and Computer Engineering courses at th... more In 2018, only 10% of students enrolled in Computer Science and Computer Engineering courses at the Federal University of Alagoas (UFAL) were women, one of the issues that highlight female underrepresentation in the academic space. In this scenario, the Katie Group appears, formed by the students of these courses, focused on initiating the process of reversing this low female representation, acting not only in academic spaces, but also in these previous educational environments, such as high school. So, with that inspiring goal in mind, the group was named after Katherine Bouman, the computer scientist responsible for the algorithm used to create the first image of a huge black
Brazilian Journal of Probability and Statistics, 2009
The Visual Computer, 2007
In this work we propose a new approach for fast visualization and exploration of virtual worlds b... more In this work we propose a new approach for fast visualization and exploration of virtual worlds based on the use of cartographic concepts and techniques. Versions of cartographic maps with different levels of details can be created by using a set of operations named cartographic generalization. Cartographic generalization employs twelve operators and domain-specific knowledge, being the contribution of this work their transposition to 3D virtual worlds. The architecture of a system for 3D generalization is proposed and the system is implemented. Differently from traditional cartographic processes, we use artificial intelligence for both selecting the key objects and applying the operators. As a case study, we present the simplification of the historical quarter of Recife (Brazil).
Abstract This paper presents the image formation process in vibro-acoustography for systems based... more Abstract This paper presents the image formation process in vibro-acoustography for systems based on sector array transducers. These transducers are an alternative to annular concave transducers. They represent an innovative technique deserving detailed ...