Lourdes Mattos Brasil

Considerations for the design of a tutoring system applied to diabetes

Diabetes Mellitus (DM) is a chronic disease affecting a large segment of worlds population. This article proposes the prototype design of a Tutoring System (TS) developed to support the educational needs related to DM treatment. The objective is to guide Type 1 patients and their families in the better routines and cautions they should follow, to improve their life quality. As the TS design is the first step of the project of an Intelligent Tutoring System design, the entire project is presented but the first step, the TS design and the pedagogical issues are described more in detail in this article.

MIDAS – Mammographic Image Database for Automated Analysis

Fabiano Fernandes1, Rodrigo Bonifacio2, Lourdes Brasil3, Renato Guadagnin4 and Janice Lamas5 1Instituto Federal de Brasilia, 2Computer Science Department, University of Brasilia, 3Post-Graduate Program in Biomedical Engineering, University of Brasilia at Gama 4Post-Graduate Program in Knowledge Management and Information Technology, Catholic University of Brasilia, 5Janice Lamas Radiology Clinic Brazil

A neuro-fuzzy-GA system architecture for helping the knowledge acquisition process

The knowledge acquisition process consists on extracting knowledge of a domain expert. This work aims to minimize the intrinsic difficulties of the knowledge acquisition process. For achieve this purpose, all possible rules from the domain expert and a set of example were obtained for a short time interval. The proposed hybrid expert system minimizes the knowledge acquisition difficulties using a new methodology. To build this hybrid architecture, several tools were used: symbolic paradigm, connectionist paradigm, fuzzy logic and genetic algorithm.

The Diagnosis Support System for Ischemic Cardiopathy: A Case Study in the Context of IACVIRTUAL Project

The Artificial Intelligence Applied in the Modeling and Implementation of a Virtual Medical Office Project (IACVIRTUAL) proposes an intelligent system to simulate a Web-based Medical Office conceived to support (i) patients who are interested in following their medical reports, (ii) professionals interested in decisions support systems for diagnosis and treatment, and (iii) students interested in learning by following available medical cases. IACVIRTUAL has several modules and among them are the Decision Support Module (DSM) and the Educational Module (EM). The Intelligent Tutor System (ITS), a sub-module of EM, is based on the MATHEMA architecture, which was conceived to provide a space for promoting cooperative interactions between an apprentice and a society of artificial tutor agents, having a human expert’s society as a backup assistance for such interactions. DSM supports medical diagnosis and subsidies EM with problem resolution using Case-Based Reasoning (CBR). CBR is an Artificial Intelligence technique that solves a new problem using adaptations of solutions of already known similar problems. The CBR model proposed for IACVIRTUAL is based on the Cross-industry Standard Process for Data Mining (CRISP-DM) and follows the 4R (Recuperation, Reutilization, Revision and Retention) cycle. Using (i) this model as fundament, (ii) a base with 1052 clinical records referring to the disease Ischemic Cardiopathy, (iii) the support of a medical expert, (iv) methods of Data Mining, and (v) the IACVIRTUAL environment; a system was developed to support medical diagnosis in the field of Cardiology. This paper intends to describe this system as well to present a case study where it was used in the context of IACVIRTUAL project.

3D Modelling of the Human Skull

Project MSE — Medical Simulation Environment—is dedicated to the creation of a propitious environment to the learning of new concepts and surgical procedures searching the connection enter the areas of accurate sciences and human beings. For this matter we present two complementary methodologies to generate three-dimensional human structures surfaces. The first methodology uses a set of digital photographs to create 3D surfaces. The second methodology uses medical scanning devices to capture the anatomy and reconstruct 3D surfaces of bones. The final 3D result can be explored in a designed WEB environment. Together these aspects make the learning and acquisition of knowledge in this environment of medical simulation easier.

FUZZYRULEXT: extraction technique of if/then rules for fuzzy neural nets

User explanation is an important function in artificial intelligence. Experience with expert systems and artificial neural networks (ANN) has shown that the ability to generate explanations is absolutely crucial for user acceptance, mainly in the medical area. ANNs traditionally have had difficulties with generating explanation structures. With this in mind, this paper presents an extraction technique of rules for a fuzzy ANN AND/OR.

A Hybrid Expert System Architecture for Medical Diagnosis

This paper deals with a new methodology for the development of an expert system (ES) using a hybrid architecture. This architecture simplifies the knowledge acquisition phase, by providing some sort of learning corresponding to the training phase of the neural network. So, it is possible to start with the nucleus of a knowledge base and the system will improve during the learning phase using examples.

Sistema Neuro-Fuzzy Para Detecção E Classificação De Microcalcificações Em Mamogramas

⎯ Breast cancer is the second cancer in the world and the most common cancer among women. Since the causes are unknown, it cannot be prevented. The presence of microcalcification clusters is an important sign for detection of breast carcinoma. Mammography is one of the most reliable exams for breast cancer detection. Early detection is the key issue for breast cancer control and computer-aided diagnosis system can help radiologists in detection and diagnosing breast abnormalities. This paper presents a fuzzy-neural system that classifies the mammogram region of interest as benign or malign. The Mini Mammographic Image Analysis Society Digital Mammogram Database (Mini-MIAS) was used to assess the system. The system accuracy achieved was in the order of 87%. Keywords⎯ Breast cancer, microcalcification, computer-aided diagnosis, Neural Networks, ANFIS. Resumo⎯ O câncer de mama é o segundo câncer mais freqüente no mundo e o tipo mais comum de câncer entre as mulheres. Como suas causas são desconhecidas, o mesmo não pode ser prevenido. A presença de microcalcificações é um sinal importante para a detecção do carcinoma mamário. A detecção precoce é um aspecto chave para o controle do câncer de mama e a mamografia é um dos exames mais confiáveis para a detecção do câncer de mama. Um sistema de diagnóstico auxiliado por computador pode ajudar os radiologistas na detecção e diagnóstico do câncer de mama, aumentando sua eficiência em até 30%. O presente estudo propõe um sistema neuro-fuzzy para a classificação da região de interesse do mamograma como maligna ou benigna. A base de dados Mini-MIAS do Mammographic Image Analysis Society Digital Mammogram Database foi utilizada para validar o modelo proposto e o mesmo atingiu uma precisão de 87%. Palavras-chave⎯ Câncer de mama, microcalcificação, sistema de diagnóstico auxiliado por computador, Redes Neurais Artificiais, ANFIS.

Hybrid Neural Solutions for Automatic Knowledge Discovery from Databases

Artificial Neural Networks (ANN) have been successfully used in a wide variety of real-world applications. However, ANN alone have not been fully employed in KDD (Knowledge Discovery in Databases) applications because they often produce incomprehensible models. Neuro-fuzzy systems and techniques for symbolic knowledge extraction have been increasingly used to represent the knowledge acquired by ANNs in a comprehensible form. This paper presents hybrid neural solutions for the KDD process, resulting from a detailed experimental investigation of three neural models (MLP, FuNN and FWD), four symbolic knowledge extraction techniques (AREFuNN, REFuNN, TREPAN and FWD) and two feature selection algorithms (FWD and the decision tree extracted by TREPAN). A large scale credit assessment application in a real-world situation was used as the test bed for the experimental investigations carried out. The results demonstrate that the benefits obtained from hybrid neural solutions are actual.

JCHAI3D - Multiplatform Framework for Healthcare Applications: Multiplatform Framework for Graphic and Haptic Processing Develop in Java

This paper presents the phases and procedures involved in the construction of a framework in order to make the process of building Virtual Reality applications, mainly for easier the health field. As study case, is presented the development of a Core Biopsy procedure simulator.