Repositorio :: Examinando por Autor "Gelvez-Almeida, E"

Examinando por Autor "Gelvez-Almeida, E"

Mostrando 1 - 20 de 20

Classification of Parkinson's disease patients based on spectrogram using local binary pattern descriptors
(IOP Publishing, 2022) Gelvez-Almeida, E; Váasquez-Coronel, A; Guatelli, R; Aubin, V; Mora, M
Extreme learning machine is an algorithm that has shown a good performance facing classi cation and regression problems. It has gained great acceptance by the scienti c community due to the simplicity of the model and its sola great generalization capacity. This work proposes the use of extreme learning machine neural networks to carry out the classi cation between Parkinson's disease patients and healthy individuals. The descriptor used corresponds to the feature vector generated applying the local binary Pattern algorithm to the grayscale spectrograms. The spectrograms are obtained from the audio signal samples from the considered repository. Experiments are conducted with single hidden layer and multilayer extreme learning machine networks comparing the results of each structure. Results show that hierarchical extreme learning machine with three hidden layers has a better general performance over multilayer extreme learning machine networks and a single hidden layer extreme learning machine. The rate of success obtained is within the ranges presented in the literature. However, the hierarchical network training time is considerably faster compared to multilayer networks of three or two hidden layers.
Estimation of the optimal number of neurons in extreme learning machine using simulated annealing and the golden section
(IOP Publishing, 2023) Gelvez-Almeida, E; Mora, M; Huérfano-Maldonado, Y; Salazar-Jurado, E; Martínez-Jeraldo, N; Lozada-Yavina, R; Baldera-Moreno, Y; Tobar, L
Extreme learning machine is a neural network algorithm widely accepted in the scientific community due to the simplicity of the model and its good results in classification and regression problems; digital image processing, medical diagnosis, and signal recognition are some applications in the field of physics addressed with these neural networks. The algorithm must be executed with an adequate number of neurons in the hidden layer to obtain good results. Identifying the appropriate number of neurons in the hidden layer is an open problem in the extreme learning machine field. The search process has a high computational cost if carried out sequentially, given the complexity of the calculations as the number of neurons increases. In this work, we use the search of the golden section and simulated annealing as heuristic methods to calculate the appropriate number of neurons in the hidden layer of an Extreme Learning Machine; for the experiments, three real databases were used for the classification problem and a synthetic database for the regression problem. The results show that the search for the appropriate number of neurons is accelerated up to 4.5× times with simulated annealing and up to 95.7× times with the golden section search compared to a sequential method in the highest-dimensional database.
Interest rates calculation in certain ordinary annuities
(IOP Publishing, 2019) Flórez, M; Vera, M; Salazar-Torres, J; Huérfano, Y; Gelvez-Almeida, E; Valbuena, O; Vera, M I; Aranguen, M
Certain annuities are annuities whose payments occur on fixed dates; while a certain ordinary annuity is one in which payments are made at the end of each established period. The calculation of the interest rate, which governs the certain ordinary annuity, involves the use of a non-analytical equation that requires the application of numerical techniques to obtain the value of the aforementioned rate. The literature indicates that any of these techniques requires one or several numerical values for initialization, which generally are estimated using trial techniques, graphical methods or values present in pre-established tables. Through this article, a new robust methodology is proposed that calculates the useful numerical values to initialize the linear interpolation technique, which is used to calculate the interest rate linked to the certain ordinary annuity. The proposed methodology generates initialization values, one by default and the other by excess, which allow us to limit the value of the certain ordinary annuity interest rate considered. Finally, we generated a new strategy that constitutes a novel mathematical model for interest rates calculation in the context of certain ordinary annuity. The percentage relative error obtained indicates the excellent performance of the aforementioned mathematical model.
Large cells cancer volumetry in chest computed tomography pulmonary images
(IOP Publishing, 2019) Huérfano, Y; Vera, M; Gelvez-Almeida, E; Vera, M I; Valbuena, O; Salazar-Torres, J
Lung cancer is the leading oncological cause of death in the world. As for carcinomas, they represent between 90% and 95% of lung cancers; among them, non-small cell lung cancer is the most common type and the large cell carcinoma, the pathology on which this research focuses, is usually detected with the computed tomography images of the thorax. These images have three big problems: noise, artifacts and low contrast. The volume of the large cell carcinoma is obtained from the segmentations of the cancerous tumor generated, in a semi-automatic way, by a computational strategy based on a combination of algorithms that, in order to address the aforementioned problems, considers median and gradient magnitude filters and an unsupervised grouping technique for generating the large cell carcinoma morphology. The results of high correlation between the semi-automatic segmentations and the manual ones, drawn up by a pulmonologist, allow us to infer the excellent performance of the proposed technique. This technique can be useful in the detection and monitoring of large cell carcinoma and if it is considering this kind of computational strategy, medical specialists can establish the clinic or surgical actions oriented to address this pulmonary pathology.
Left atrial appendage automatic segmentation, in computed tomography images
(IOP Publishing, 2019) Huérfano, Y; Vera, M; Vera, M I; Valbuena, O; Gelvez-Almeida, E; Salazar-Torres, J; Molina, V
The left atrial appendage is one of the anatomical places where most frequently blood thrombi occur. When migrating from the appendage, these thrombi, become blood emboli that, potentially, can compromise the physiology and/or anatomy of cardiac or cerebral blood vessels, being able to generate cerebrovascular events. The left atrial appendage segmentation is very difficult due, mainly, to its location and the identical densitometric information presents into of this appendage and around of the left atrium. In this paper, an automatic technique is proposed to segment this appendage with the purpose of generating important information to the procedure called left atrial appendage surgical closure. This information is linked to the volume and the diameters of the left atrial appendage. The technique consists of a digital pre-processing stage, based on filtering processes and definition of a region of interest and, of one segmentation stage that considers a clustering method. The results are promising and they allow us to calculate useful quantitative variables when characterizing the most lethal appendix of the human body represented by the mentioned appendage. These results are very important in clinical processes where both the shape and volume of this appendage are vital for detecting and monitoring some vascular diseases such as cardiac embolism, arterial hypertension and stroke, among others.
Mathematical argumentation in the classroom
(IOP Publishing, 2019) Salazar-Torres, J; Vera, M; Contreras, Y; Gelvez-Almeida, E; Valbuena, O; Barrera, D; Rincón, O
The article shares some elements of comprehensive type about "mathematical argumentation in the classroom"; whose analysis, was made from two fundamental categories in the development of an oral mathematical argumentation process for the conviction, contradiction and validation of a written mathematical argumentation process. The research addressed two central categories of argumentation as a discursive form, the first one is the epistemic position, and the second one is the discursive position that students unveil at the time of mathematically arguing the solution to a problem situation. The research was developed under the interpretative paradigm through the design of a case study directed by the theory and technique of a focal group, for the collection of information. In the findings, difficulties in the passage were evidenced from the semantic to the theoretical from the epistemic position; regarding the discursive position, the presence of three discursive forms was revealed: description, explanation and argumentation, the latter being the least used by the students.
Newton-Raphson method initialization for non-analytical equations solution linked to anticipated annuities
(IOP Publishing, 2019) Vera, M; Flórez, M; Salazar-Torres, J; Huérfano, Y; Gelvez-Almeida, E; Valbuena, O; Vera, M I; Aranguen, M
The series of payments made in equal intervals of time is known, in the world of financial mathematics, as an annuity. An anticipated annuity is one whose periodic payment expires at the beginning of the established payment interval. The non-analytical equation that allows us to calculate the interest rate, linked to the anticipated annuity, can be solved using several numerical methods, in particular, the numerical method called Newton-Rhapson. The main problem with this method is its initialization, which requires of one starting point that, usually, is estimated without any scientific background or using random or arbitraries mechanisms. In order to address this problem, in this paper, we establish as main objective to demonstrate that the Newton-Rhapson method can be initialized using only the data, of an anticipated annuity, identified as capital, income and payment intervals without the need to use the initialization strategies, reported in the literature. Through this article, a strategy is presented that allow us to calculate the value of the AA interest rate using the MNR. The value of the error generated for the problematic considered in order to assess the quality of the work performed, is a clear indicator of the good performance of the proposed strategy. This strategy for obtaining the starting point of the aforementioned numerical method is useful in the financial mathematical context, for example, when is necessary the interest rate calculation.
Parallel methods for linear systems solution in extreme learning machines: an overview
(2020) Gelvez-Almeida, E; Baldera-Moreno, Y; Huérfano, Y; Vera, M; Mora, M; Barrientos, R
This paper aims to present an updated review of parallel algorithms for solving square and rectangular single and double precision matrix linear systems using multi-core central processing units and graphic processing units. A brief description of the methods for the solution of linear systems based on operations, factorization and iterations was made. The methodology implemented, in this article, is a documentary and it was based on the review of about 17 papers reported in the literature during the last five years (2016-2020). The disclosed findings demonstrate the potential of parallelism to significantly decrease extreme learning machines training times for problems with large amounts of data given the calculation of the Moore Penrose pseudo inverse. The implementation of parallel algorithms in the calculation of the pseudo-inverse will allow to contribute significantly in the applications of diversifying areas, since it can accelerate the training time of the extreme learning machines with optimal results.
Problem solving strategy in the teaching and learning processes of quantitative reasoning
(IOP Publishing, 2019) Barrera, D; Salazar-Torres, J; Vera, M; Gelvez-Almeida, E
The study presents an analysis of Polya's problem-solving strategy used in the training processes of quantitative reasoning competence in students of the Universidad Simón Bolívar, San José de Cúcuta, Colombia. The research was based on a descriptive design and had an intentional sample of 58 students who were studying the sciences and general competencies elective. For the collection of information, a diagnostic test (pre-test) and a final test (post-test) were applied, in order to check the incidence of the applied strategy. The results showed a significant improvement in the final results obtained by the students in each of the processes formed: interpretation, representation and modeling, and argumentation.
Pulmonary adenocarcinoma characterization using computed tomography images
(IOP Publishing, 2019) Huérfano, Y; Vera, M; Valbuena, O; Gelvez-Almeida, E; Salazar-Torres, J
Lung cancer is one of the pathologies that sensitively affects the health of human beings. Particularly, the pathology called pulmonary adenocarcinoma represents 25% of all lung cancers. In this research, we propose a semiautomatic technique for the characterization of a tumor (adenocarcinoma type), present in a three-dimensional pulmonary computed tomography dataset. Following the basic scheme of digital image processing, first, a bank of smoothing filters and edge detectors is applied allowing the adequate preprocessing over the dataset images. Then, clustering methods are used for obtaining the tumor morphology. The relative percentage error and the accuracy rate were the metrics considered to determine the performance of the proposed technique. The values obtained from the metrics used reflect an excellent correlation between the morphology of the tumor, generated manually by a pneumologist and the values obtained by the proposed technique. In the clinical and surgical contexts, the characterization of the detected lung tumor is made in terms of volume occupied by the tumor and it allows the monitoring of this disease as well as the activation of the respective protocols for its approach.
Renal lithiasis detection in uro-computed tomography using a non-parametric technique
(IOP Publishing, 2019) Rodríguez-Ibáñez, R; Vera, M I; Vera, M; Gelvez-Almeida, E; Huérfano, Y; Valbuena, O; Salazar-Torres, J
Renal lithiasis is the pathology that causes nephritic colic, which is one of the most frequent reasons for consultation in emergency medical services. According to the size, location, hardness and number of stones present in the urinary system, usually in the human kidney, it is established to which form of treatment is suitable for the patient. These kidney stones can be analyzed by means of biopsy or imaging modalities such as computed tomography images. This type of images has challenging problems called noise, artifacts and low contrast. In this paper, in order to address these problems, a non-parametric semi-automatic computational technique is developed for detecting kidney stones, present in computed tomography images, using digital image processing techniques based on a smoothing filter and an edge detector. Finally, the size and position of the stones present in the images are calculated and a precision metric is considered to compare the manual segmentation, performed by an urologist, and the one generated by the NPCT, obtaining an excellent correlation. This technique can be useful in the renal lithiasis detection and if it is considering this kind of computational strategy, medical specialists can establish the clinic or surgical actions oriented to address this pathology.
The rubric as an assessment strategy in the mathematical argumentation process
(IOP Publishing, 2020) Salazar-Torres, J; Vera, M; Contreras, Y; Gelvez-Almeida, E; Huérfano, Y; Valbuena, O
The article shares the proposal of an analytical rubric as a strategy for the assessment and monitoring of learning outcomes in students who develop an argumentative plot from the mathematics field, to solve any problem situation in daily life. The study was based on the theory of mathematical argumentation proposed by Duval and the contributions of León and Calderón, as well as the dimensions presented to us by the logical frameworks in the design of analytical rubrics. The research was developed under the social critical paradigm through the design of pedagogical action research, and the focus group technique was used for the collection of information composed by five professors from the department of basic sciences. As a result, a collective rubric that, in addition to generating processes of self-assessment and self-training in teachers, evidences a decrease in the existent subjectivity of the evaluation processes, thus strengthening its objectivity.
Semi-automatic detection of hepatic tumor in computed tomography images
(IOP Publishing, 2019) Sáenz, F; Vera, M; López, J; Huérfano, Y; Valbuena, O; Vera, M I; Gelvez-Almeida, E; Salazar-Torres, J
In this work, the main purpose is develop a computational segmentation strategy for liver tumor semiautomatic detection. This strategy considers three-dimensional computed tomography images and it consists of techniques application that, on the one hand, diminish the noise and detect the edges of the objects present in those images and, on the other hand, generate the liver tumor morphology. For this, the sequence of techniques composed of gaussian smoothing, gradient magnitude, median filter, region growing and binary morphological dilation are used. The value obtained, for the metric called Dice score, show a good correlation between manual segmentation, performed by a hepatologist, and the tumor segmentation obtained using the proposed technique. This type of segmentation is the extreme utility for the characterization of hepatic tumors and the planning of the clinical behavior to be followed in the treatment of this human liver disease.
Semi-automatic detection of the evolutionary forms of visceral leishmaniasis in microscopic blood smears
(IOP Publishing, 2019) Salazar, J; Vera, M; Huérfano, Y; Vera, M I; Gelvez-Almeida, E; Valbuena, O
Leishmaniasis is a complex group of diseases caused by obligate unicellular and intracellular eukaryotic protozoa of the leishmania genus. Leishmania species generate diverse syndromes ranging from skin ulcers of spontaneous resolution to fatal visceral disease. These syndromes belong to three categories: visceral leishmaniasis, cutaneous leishmaniasis and mucosal leishmaniasis. The visceral leishmaniasis is based on the reticuloendothelial system producing hepatomegaly, splenomegaly and lymphadenopathy. In the present article, a semiautomatic segmentation strategy is proposed to obtain the segmentations of the evolutionary shapes of visceral leishmaniasis called parasites, specifically of the type amastigote and promastigote. For this purpose, the optical microscopy images containing said evolutionary shapes, which are generated from a blood smear, are subjected to a process of transformation of the color intensity space into a space of intensity in gray levels that facilitate their subsequent preprocessing and adaptation. In the preprocessing stage, smoothing filters and edge detectors are used to enhance the optical microscopy images. In a complementary way, a segmentation technique that groups the pixels corresponding to each one of the parasites, presents in the considered images, is applied. The results reveal a high correspondence between the available manual segmentations and the semi-automatic segmentations which are useful for the characterization of the parasites. The obtained segmentations let us to calculate areas and perimeters associated with the parasites segmented. These results are very important in clinical context where both the area and perimeter calculated are vital for monitoring the development of visceral leishmaniasis.
Smart operator for the human liver automatic segmentation, present in medical images
(IOP Publishing, 2019) Vera, M; Sáenz, F; Huérfano, Y; Gelvez-Almeida, E; Vera, M I; Salazar-Torres, J; Valbuena, O
The segmentation of the human body organ called liver is a highly challenging problem due to the noise, artifacts and the low contrast exhibited by the anatomical structures located around the liver and that are present in digital images, generated by any modality of medical images. The main modalities are: ultrasound, nuclear emission, magnetic resonance and the gold standard called multi-slice computed tomography. In this paper, with the objective of to address this problem, we consider multi-slice computed tomography images and we propose an automatic strategy based on two phases. In the first phase, a digital filtering bank is used for diminishing the noise effect and the artifacts impact in the quality of images. In the second phase, called liver detection, we use a smart operator based on least squares support vector machines for generating both the morphology and the volume of liver. The application of this strategy allows generating the morphology of the liver in a precise and efficient manner as it was demonstrated by the metrics used to assess its performance. These results are very important in clinical-surgical processes where both the shape and volume of liver are vital for monitoring some liver diseases that can affect the normal liver physiology.
Space-occupying lesions identification in mammary glands using a mixed computational strategy
(IOP Publishing, 2019) Vargas, S; Vera, M I; Vera, M; Salazar-Torres, J; Huérfano, Y; Valbuena, O; Gelvez-Almeida, E
Abstract. The mammary pathology can manifest itself in multiple ways and originates spaceoccupying lesions. The breast cancer is a space-occupying lesion, which is highly prevalent, especially in women, and worldwide it is one of the leading causes of morbidity and mortality in this population. The main image modality for breast cancer detection is the magnetic resonance but this kind of image modality introduces several imperfections that affect the image quality. Some of these imperfections or problems are: inhomogeneity in the anatomical structures, riccian noise and artifacts. These problems make the analysis of the image information a real challenge. To address these problems, in this paper, we propose a computational technique able to extract a space-occupying lesion linked to breast cancer, present in magnetic resonance images. For this, the original image is processed with statisticalarithmetic filters and segmentation algorithms based on thresholding and multi-seed region growing techniques. The results, based on Dice score, show that the proposed technique is suitable for segmenting the breast cancer due high correlation between semi-automatic and manual segmentations. This technique can be useful in the detection, characterization and monitoring of this type of cancer and it can let to medical doctors to realize their work more efficiently.
Use of computational realistic models for the cardiac ejection fraction calculation
(IOP Publishing, 2019) Huérfano, Y; Vera, M; Vera, M I; Valbuena, O; Gelvez-Almeida, E; Salazar-Torres, J
Ejection fraction is one of the most useful clinical descriptors to determine the cardiac function of a subject. For this reason, obtaining the value of this descriptor is of vital importance and requires high precision. However, in the clinical routine, to generate the mentioned descriptor value, a geometric hypothesis is assumed, obtaining an approximate value for this fraction, usually by excess, and which is a dependent-operator. The aim of the present work is to propose the accurate calculation of the ejection fraction from realistic models, obtained computationally, of the cardiac chamber called right ventricle. Normally, the geometric hypothesis that makes this ventricle coincide with a pyramidal type geometric shape, is not usually, fulfilled in subjects affected by several cardiac pathologies, so as an alternative to this problem, the computational segmentation process is used to generate the morphology of the right ventricle and from it proceeds to obtain, accurately, the ejection fraction value. In this sense, an automatic strategy based on no-lineal filters, smart operator and region growing technique is propose in order to generate the right ventricle ejection fraction. The results are promising due we obtained an excellent correspondence between the manual segmentation and the automatic one generated by the realistic models.
Usefulness of cutting planes in the hierarchical segmentation of cardiac anatomical structures
(IOP Publishing, 2019) Vera, M; Valbuena, O; Huérfano, Y; Vera, M I; Gelvez-Almeida, E; Salazar-Torres, J
A spatial geometric plane is defined by the three-dimensional coordinates of a pair of spatial points and the direction that the normal vector establishes, which is formed by joining those points by means of an oriented line segment. This type of planes, in three-dimensional images, is extremely useful as an alternative solution to the problem of low contrast that exhibit the anatomical structures present in cardiac computed tomography images. To do this, after using a predetermined filter bank and in order to define a region of interest, a smart operator based on least squares support vector machines is trained and validated in order to detect the aforementioned coordinates which enables the location of the plane, in the three-dimensional space that contains the considered images. Once the structure that is required to segment is identified, a discriminant function is used that cancels all information not linked to this structure. In this work, the segmentation of the left ventricle, based on region growing technique, is firstly considered and then the left atrium is segmented considering region growing technique and an inverse discriminant function. The results show an excellent correspondence relationship when the spatial union of both structures is made.
Usefulness of digital images segmentation in pulmonary transplantation
(IOP Publishing, 2019) Gelvez-Almeida, E; Huérfano, Y; Vera, M; Vera, M I; Valbuena, O; Salazar-Torres, J
In the presence of pulmonary pathologies such as chronic obstructive pulmonary disease, diffuse pulmonary disease and cystic fibrosis, among others, it is common to require the removal or replacement of a portion of lungs. There are several requirements for both donors and organ receivers (recipients) established in the literature. May be the main one is the volume that the donor's lungs occupy in the thoracic cavity. This parameter is vital because if the volume of the lungs exceeds the thoracic cavity of the recipients the transplant, logically, is unfeasible for physical reasons such as the incompatibility between the receiver lung volume and the donor lung volume. In this sense, the present paper proposes the creation of a hybrid technique, based on digital image processing techniques application to raise the quality of the information related to lungs captured in three-dimensional sequences of computed tomography and for generating the morphology and the volumes of the lungs, belonging to a patient. During the filtering stage median, saturated and gradient magnitude filters are applied with the purpose of addressing the noise and artefacts images problems; whereas during the segmentation stage, methods based on clustering processes are used to extract the lungs from the images. The values obtained for the metric that assesses the quality of the hybrid computational technique reflect its good performance. Additionally, these results are very important in clinical processes where both the shapes and volumes of lungs are vital for monitoring some lung diseases that can affect the normal lung physiology.
Volumetric quantification in ovarian pathology using abdomino-pelvic computed tomography
(IOP Publishing, 2019) Valbuena, O; Vera, M; Vera, M I; Gelvez-Almeida, E; Huérfano, Y; Borrero, M; Salazar-Torres, J; Salazar, W
Pathological ovary is categorized into cystic tumors, solid tumors and mixed, according to the content of the affected ovary. Accordingly, the degree of benignity or malignity thereof is established. The imaging study for the preliminary morphological assessment of PO is ultrasound, in its pelvic and transvaginal modalities, for which wellestablished criteria are available. Once the ultrasound findings suggest malignancy, complementary studies such as abdominal-pelvic tomography images and tumor markers are requested. This type of images has challenging problems called noise, artifacts and low contrast. In this paper, in order to address these problems, a computational technique is proposed to characterize a pathological ovary. To do this, a thresholding and the median and gradient magnitude filters are applied, preliminarily, to complete the preprocessing stage. Then, during the segmentation, the algorithm of region growing is used to extract the threedimensional morphology of the pathological ovary. Using this morphology, the volume of the pathological ovary is calculated and it allows selecting the surgical-medical behavior to approach this kind of ovary. The validation of the proposed technique indicates that the results are promising. This technique can be useful in the detection and monitoring the diseases linked to pathological ovary.

Examinando por Autor "Gelvez-Almeida, E"

Resultados por página

Opciones de ordenación