Fabrízzio Alphonsus A. M. N. Soares

Comparative Performance Analysis of Machine Learning Classifiers in Detection of Childhood Pneumonia Using Chest Radiographs

Abstract This work extends PneumoCAD, a Computer-Aided Diagnosis system for detecting pneumonia in infants using radiographic images [1] , with the aim of improving the systems accuracy and robustness. We implement and compare three contemporary machine learning classifiers, namely: Naive Bayes, K-Nearest Neighbor (KNN), and Support Vector Machines (SVM). Results of our experiments demonstrate that the SVM classifier produces the best overall results.

Method for Text Entry in Smartwatches Using Continuous Gesture Recognition

This work proposes a method that allows the entry of text in smartwatches using gestures based on geometric forms. For this it is proposed the development of a prototype capable of inserting a letter with no more than two user interactions. Gesture recognition is performed using the incremental recognition algorithm. A set of gestures with lines and curves were created to be recognized by the incremental recognition algorithm, generated from the reduced equation of the line and the reduced equation of the circumference, respectively. After recognizing the gestures, they are sent to a classifier Naïve Bayes which is responsible for predicting the letter that will be inserted. The Naïve Bayes classifier was trained with a user gesture base that drew all the letters of the alphabet using only the gestures available in the set presented to them. Using the gesture base and the classifier Naïve Bayes a prototype was developed for smartwatches that automatically suggests the most likely letters to be inserted. The prototype was used to perform an experiment, during the experiment the users inserted the five most frequent letters and the five less frequent letters of the English language. The results of the experiment show that the prototype is able to recognize a letter with at most two interactions between the user and the smartwatch. The analysis of the usability and experience test shows that the prototype has generalized potential for use, since it allows the entry of text with up to two interactions and with a 100% hit rate for the most frequent letters and 95,14% For less frequent letters.

BrailleÉcran: A Braille Approach to Text Entry on Smartphones

Nowadays, due to touchscreen, mobile devices have become much more dynamic than in the past. However, due the same reason, those devices are less accessible for blind people. It is expected that in 2018, over half of mobile phone users will have a smartphone, therefore researches about mobile accessibility are very important. So, the aim of this systematic review is to cooperate with new research about methods for braille text entry on smartphones. The systematic searches in 5 databases, resulted in 11 papers that answered the research questions that grounded this work.

Braille Text Entry on Smartphones: A Systematic Review of the Literature

Nowadays, due to touchscreen, mobile devices have become much more dynamic than in the past. However, due the same reason, those devices are less accessible for blind people. It is expected that in 2018, over half of mobile phone users will have a smartphone, therefore researches about mobile accessibility are very important. So, the aim of this systematic review is to cooperate with new research about methods for braille text entry on smartphones. The systematic searches in 5 databases, resulted in 11 papers that answered the research questions that grounded this work.

Recursive diameter prediction for calculating merchantable volume of eucalyptus clones using Multilayer Perceptron

A very common problem in forestry is the realization of the forest inventory. The forest inventory is very important because it allows the trading of medium- and long-term timber to be extracted. On completion , the inventory is necessary to measure different diameters and total height to calculate their volumes. However, due to the high number of trees and their heights, these measurements are an extremely time consuming and expensive. In this work, a new approach to predict recursively diameters of eucalyptus trees by means of Multilayer Perceptron artificial neural networks is presented. By taking only three diameter measures at the base of the tree, diameters are predicted recursively until they reach the value of 4 cm, with no previous knowledge of total tree height. The training was conducted with only 10% of the total trees planted site, and the remaining 90% of total trees were used for testing. The Smalian method was used with the predicted diameters to calculate merchantable tree volumes. To check the performance of the model, all experiments were compared with the least square polynomial approximator and the diameters and volumes estimates with both methods were compared with the actual values measured. The performance of the proposed model was satisfactory when predicted diameters and volumes are compared to actual ones.

Recursive diameter prediction for calculating merchantable volume of Eucalyptus clones without previous knowledge of total tree height using artificial neural networks

In this work, diameters of Eucalyptus trees are predicted by means of Multilayer Perceptron and Radial Basis Function artificial neural networks. By taking only three diameter measures at the base of the tree, diameters are predicted recursively until they reach the value of minimum merchantable diameter, with no previous knowledge of total tree height. It was considered the diameter top of 4cm outside bark as minimum merchantable diameter. The training was conducted with only 10% of the trees from the total planted site. The Smalian method utilizes the predicted diameters to calculate merchantable tree volumes. The performance of the proposed model was satisfactory when predicted diameters and volumes are compared to actual ones.

Text input in Smartwatches Based Gestures Using Geometric Shape

This paper proposes a method for text input based on gestures to be used in smartwatches using geometric shapes. To make the recognition of gestures, we used the incremental recognition algorithm gestures. a template with straight curves were developed using the reduced equation of the circle. Using this template, thirty users have entered all the letters of the alphabet three times each in three groups totaling ninety inserts for each letter. Gestures entered by users have been used to train a Naive Bayes classifier that calculates the probability of insertion for each letter to from the user-entered gestures. During the development work was also carried out a study of the most frequent letters of the Portuguese language. Another partial result of the project is a prototype in which the user enters all the letters of the alphabet using the template gestures. By the time the user enters a gesture prototype automatically suggests the most frequent letters using the Naïve Bayes classifier.

Wrist Player: A Smartwatch Gesture Controller for Smart TVs

Emerging technology on mobile and wearable market, smartwatches have embedded movement sensors whose potential is yet to be fully explored. This paper proposes an interaction method with smart TVs via gestures performed by persons wrist using a smartwatch. Detailed architecture and implementation for a complete prototype, named Wrist Player, is presented. A user study is also conducted, in order to evaluate the prototype performance and the users interest on the proposal. Results show that the method works very well, with participants reporting having a good experience with the prototype. We present our insights on the concept, challenges faced in our research and ideas for future studies.

Algoritmo Genético Multi-Objetivo Para Seleção de Variáveis em Problemas de Classificação Multivariada: Um Estudo de Caso na Verificação de Adulteração de Biodiesel

Lucas de Almeida Ribeiro Anderson da Silva Soares Instituto de Informática Universidade Federal de Goiás Goiânia, Goiás Email: (lucasaribeiro, anderson)@inf.ufg.br Clarimar José Coelho Departamento de Ciência da Computação Pontifı́cia Universidade Católica de Goiás Goiânia, Goiás Fabrı́zzio A. A. Melo Nunes Soares Telma Woerle de Lima Carlos Antônio Campos Jorge Instituto de Informática Universidade Federal de Goiás Goiânia, Goiás

A New Approach to Detect Use of Alcohol Through Iris Videos Using Computer Vision

In all modern society the increase in alcohol consumption has caused many problems and the potential harmful effects of alcohol on human health are known. There are some ways to identify alcohol in a person, but they are invasive and embarrassing for people. This work proposes a new non-invasive and simple test to detect use of alcohol through of pupillary reflex analysis. The initial results present rates near 85% in the correct identification using algorithms for pattern recognition, demonstrating the efficacy of the test method.