Reconocimiento de rutas biosintéticas para semioquímicos mediante técnicas de aprendizaje de máquina

En este trabajo consideramos 148 semioquímicos reportados para la familia Scarabaeidae, cuya estructura química fue caracterizada empleando un conjunto de 200 descriptores moleculares de cinco clases distintas. La selección de los descriptores más discriminantes se realizó con tres técnicas: análisis de componentes principales, por cada clase de descriptores, bosques aleatorios y Boruta-Shap, aplicados al total de descriptores. A pesar de que las tres técnicas son conceptualmente diferentes, seleccionan un número de descriptores similar de cada clase. Propusimos una combinación de técnicas de aprendizaje de máquina para buscar un patrón estructural en el conjunto de semioquímicos y posteriormente realizar la clasificación de estos. El patrón se estableció a partir de la alta pertenencia de un subconjunto de estos metabolitos a los grupos que fueron obtenidos por un método de agrupamiento basado en lógica difusa, C-means; el patrón descubierto corresponde a las rutas biosintéticas por las cuales se obtienen biológicamente. Esta primera clasificación se corroboró con el empleo de mapas autoorganizados de Kohonen. Para clasificar aquellos semioquímicos cuya pertenencia a una ruta no quedaba claramente definida, construimos dos modelos de perceptrones multicapa, los cuales tuvieron un desempeño aceptable.

In this work we consider 148 semiochemicals reported for the family Scarabaeidae, whose chemical structure was characterized using a set of 200 molecular descriptors from five different classes. The selection of the most discriminating descriptors was carried out with three different techniques: Principal Component Analysis, for each class of descriptors, Random Forests and Boruta-Shap, applied to the total of descriptors. Although the three techniques are conceptually different, they select a similar number of descriptors from each class. We proposed a combination of machine learning techniques to search for a structural pattern in the set of semiochemicals and then perform their classification. The pattern was established from the high belonging of a subset of these metabolites to the groups that were obtained by a grouping method based on fuzzy C-means logic; the discovered pattern corresponds to the biosynthetic pathway by which they are obtained biologically. This first classification was corroborated with Kohonen's self-organizing maps. To classify those semiochemicals whose belonging to a biosynthetic pathway was not clearly defined, we built two models of Multilayer Perceptrons which had an acceptable performance.

Neste trabalho consideramos 148 semioquímicos reportados para a família Scarabaeidae, cuja estrutura química foi caracterizada usando um conjunto de 200 descritores moleculares de 5 classes diferentes. A seleção dos descritores mais discriminantes foi realizada com três técnicas diferentes: Análise de Componentes Principais, para cada classe de descritores, Florestas Aleatórias e Boruta-Shap, aplicadas a todos os descritores. Embora as três técnicas sejam conceitualmente diferentes, elas selecionaram um número semelhante de descritores de cada classe. Nós propusemos uma combinação de técnicas de aprendizado de máquina para buscar um padrão estrutural no conjunto de semioquímicos e então realizar sua classificação. O padrão foi estabelecido a partir da alta pertinência de um subconjunto desses metabólitos aos grupos que foram obtidos por um método de agrupamento baseado em lógica fuzzy, C-means; o padrão descoberto corresponde às rotas biossintéticas pelas quais eles são obtidos biologicamente. Essa primeira classificação foi corroborada com o uso dos mapas auto-organizados de Kohonen. Para classificar os semioquímicos cuja pertença a uma rota não foi claramente definida, construímos dois modelos de Perceptrons Multicamadas que tiveram um desempenho aceitável.

Lung nodule segmentation based on fuzzy c-means clustering and improved random walk algorithm / 生物医学工程学杂志

Accurate segmentation of pulmonary nodules is an important basis for doctors to determine lung cancer. Aiming at the problem of incorrect segmentation of pulmonary nodules, especially the problem that it is difficult to separate adhesive pulmonary nodules connected with chest wall or blood vessels, an improved random walk method is proposed to segment difficult pulmonary nodules accurately in this paper. The innovation of this paper is to introduce geodesic distance to redefine the weights in random walk combining the coordinates of the nodes and seed points in the image with the space distance. The improved algorithm is used to achieve the accurate segmentation of pulmonary nodules. The computed tomography (CT) images of 17 patients with different types of pulmonary nodules were selected for segmentation experiments. The experimental results are compared with the traditional random walk method and those of several literatures. Experiments show that the proposed method has good accuracy in the segmentation of pulmonary nodule, and the accuracy can reach more than 88% with segmentation time is less than 4 seconds. The results could be used to assist doctors in the diagnosis of benign and malignant pulmonary nodules and improve clinical efficiency.

An Improved Classification System for Brain Tumours Using Wavelet Transform and Neural Network / Un sistema de clasificación mejorado para los tumores cerebrales usando la transformada de ondeletas y la red neuronal

ABSTRACT This paper presents an improved classification system for brain tumours using wavelet transform and neural network. The anisotropic diffusion filter was used for image denoising, and the performance of the oriented rician noise reducing anisotropic diffusion (ORNRAD) filter was validated. The segmentation of the denoised image was carried out by fuzzy c-means clustering. The features were extracted using symlet and coiflet wavelet transforms, and the Levenberg-Marquardt algorithm based neural network was used to classify the magnetic resonance (MR) images. This classification technique of MR images was tested and analysed with existing methods, and its performance was found to be satisfactory with a classification accuracy of 93.24%. The developed system could assist physicians in classifying MR images for better decision-making.

RESUMEN Este artículo presenta un sistema de clasificación mejorado para los tumores de cerebro usando la transformada de ondeletas (transformada wavelet) y la red neuronal. El filtro de difusión anisotrópica fue utilizado para la eliminación del ruido de la imagen, y se validó el funcionamiento del filtro de difusión anisotrópica orientado a reducir el ruido riciano (ORNRAD, siglas en inglés). La segmentación de la imagen 'desruidizada ' (denoised) fue realizada mediante el agrupamiento difuso c-means fuzzy. Las características fueron extraídas usando las transformadas de ondeletas symlet y coiflet, y la red neuronal basada en el algoritmo de Levenberg-Marquardt fue utilizada para clasificar las imágenes de resonancia magnética (RM) imágenes. Esta técnica de clasificación de imágenes de RM fue probada y analizada con métodos existentes, y se halló que su rendimiento era satisfactorio con una precisión de clasificación de 93.24%. El sistema desarrollado podría ayudar a los médicos a clasificar imágenes de RM para una mejor toma de decisiones.

Implementación de un sistema de imagenología infrarroja para la detección vascular del antebrazo y mano / Implementation of an infrared imaging system for vascular detection in the forearm and hand

RESUMEN: En el área clínica son procedimientos comunes la venopunción, la colocación de catéteres, inyecciones intravenosas, etc. La visualización de las venas en algunas personas es compleja. En este trabajo se presenta el desarrollo de un sistema capaz de realzar la distribución de las venas en el antebrazo de una persona para, en un trabajo futuro, ayudar de forma no invasiva a localizar las venas en los procedimientos de venopunción. Para llevar a cabo el desempeño de esta tarea se utilizó una cámara web, a la cual se le ha extraído el filtro que impide el paso de luz infrarroja y es sustituido por otro que lo permite. Para mejorar la detección de las venas se le agregó a la cámara un arreglo de LEDs emisores de luz infrarroja (830nm). Las imágenes obtenidas fueron procesadas mediante la ecualización del histograma adaptable y clasificadas por dos métodos, el primero basado en el algoritmo Fuzzy C-Means, el segundo basado en un modelo probabilístico de tipo Bayes, técnicas del área de inteligencia artificial, presentadas como alternativa en el procesamiento de imágenes. Para la obtención de las imágenes se consideraron las regiones anteriores y exteriores del antebrazo izquierdo y derecho de cada sujeto generando una base de datos. Este sistema también tiene aplicación en la detección de venas varicosas debido a que se puede dar un seguimiento a la dilatación de las venas.

ABSTRACT: The venipuncture, the catheterization and intravenous (IV) injections are some of the common procedures in the clinical practice. The location of the veins may be complex in some patients. In this paper a system able to enhance the vein distribution in a patient's forearm in order to help, in future works, to locate the veins in a non-invasive way and accomplish the IV procedures, is described. To carry out this work a web cam was used, the filter that blocks out the infrared light has been removed and replaced for one who does not. To increase the vein detection an array of infrared LEDs (830 nm) was attached. The resulting images were processed using the adaptive histogram equalization and then classified by two methods, the first one based on the Fuzzy C-Means Algorithm, and the second based in a Bayesian probabilistic model. For the image acquisition, the anterior-exterior regions of the left and right forearm of each subject were considered to generate a data base. This system also has relevance in the detection of varicose veins since is able to monitor the vein dilatation.

Study on Mongolian Medicine Prescription Classification Method Based on Fuzzy C-means Algorithm / 中国中医药信息杂志

Objective To classify Mongolian medicine prescription by using fuzzy c-means algorithm (FCM) and hard c-means algorithm (HCM); To explore the rationality of two kinds of clustering algorithm. Methods 27 Mongolian medicine prescriptions for treating Heiyi disease from Chuan Tong Meng Yao Yu Fang Ji were set as experimental data, and the data were preprocessed first. MS Visual Studio 2010 platform was used, and C# language was used for research and development. Chinese version and Mogolian version were implemented with WindowFrom and WPF technology, respectively. The medicine prescriptions were classified into 3, 4, 5, and 6 types by using FCM and HCM. Results All categorization with zero classification showed the existence of inclusion phenomena. The medicine in the classification results obtained by the two kinds of clustering algorithm did not exist cross. FCM could produce clustering results with smaller quantity difference and the more uniform classification compared with HCM. Conclusion The two algorithms are correct and reasonable, in which FCM algorithm has better clustering effect, and can be widely used in Mongolian prescription analysis, with a purpose to provide data supports for the research and development of new medicine.

A Modified Kernel-based Fuzzy C-Means Algorithm For Images Segmentation / 医疗卫生装备

Objective To segment brain magnetic resonance (MR) images corrupted by noises. Methods We presented a novel Fuzzy C-Means (FCM) algorithm for image segmentation. The algorithm was by modifying the objective function in the conventional FCM. Firstly,by using kernel method,the original Euclidean distance in the FCM was replaced by a kernel-induced distance. Then,a spatial penalty term was added to the objective function to compensate the influence of the neighboring pixels on the center pixel. Results Segmentation results on a four-class synthetic image corrupted by salt & pepper noise shows that the new algorithm is less speckled and smoother. The new algorithm is applied to simulation MR images and is shown to have less misclassification rate than the other FCM-based methods. Conclusion The results of experiments show that the proposed algorithm is more robust to noise than other FCM-based methods.

Method for obtaining prior information of electrical impedance tomography / 医疗卫生装备

The reverse problem of Electrical impedance tomography(EIT) is a highly ill-posed problem.It is concluded that spatial prior information could improve the final image quality.This paper proposes a new method for obtaining prior information.By this method,the inspected cross-section contour and internal structure for EIT can be achieved.