Search

International audience; This study investigates the processing of voice signals for detecting Parkinson's disease. This disease is one of the neurological disorders that affect people in the world most. The approach evaluates the use of eighteen feature extraction techniques and four machine learning methods to classify data obtained from sustained phonation and speech tasks. Phonation relates to the vowel /a/ voicing task and speech to the pronunciation of a short sentence in Lithuanian language. The audio tasks were recorded using two microphone channels from acoustic cardioid (AC) and a smartphone (SP), thus allowing to evaluate the performance for different types of microphones. Five metrics were employed to analyze the classification performance: Equal Error Rate (EER) and Area Under Curve (AUC) measures from Detection Error Tradeoff (DET) and Receiver Operating Characteristic curves, Accuracy, Specificity, and Sensitivity. We compare this approach with other approaches that use the same data set. We show that the task of phonation was more efficient than speech tasks in the detection of disease. The best performance for the AC channel achieved an accuracy of 94.55%, AUC 0.87, and EER 19.01%. When using the SP channel, we have achieved an accuracy of 92.94%, AUC 0.92, and EER 14.15%.

Novel segmentation methods based on models of deformable active contours are constantly proposed and validated in different fields of knowledge, with the aim to make the detection of the regions of interest standard. This paper propose a new method called Optimum Path Snakes (OPS), a new adaptive algorithm and free of parameters to define the total energy of a active contour model with automatic initialization and stop criteria. In the experimental assessment, the OPS is compared against some approaches commonly used in the following fields, such as vector field convolution, gradient vector flow, and other specialists methods for lung segmentation using thorax computed tomography images. The segmentation of regions with stroke was carried out with methods based on region growing, watershed and a specialist level set approach. Statistical validations metrics using Dice coefficient (DC) and Hausdorff distance (HD) were also evaluated, as well as the processing time. The results showed that the OPS is a promising tool for image segmentation, presenting satisfactory results for DC and HD, and, many times, superior to the other algorithms it was compared with, including those generated by specialists. Another advantage of the OPS is that it is not restricted to specific types of images, neither applications.

The eagerness and necessity to develop so-called smart applications has taken the Internet of Things (IoT) to a whole new level. Industry has been implementing services that use IoT to increase productivity as well as management systems over the past couple of years. Such services are now encroaching on wind energy, which nowadays is the most acceptable source among renewable energies for electricity generation. This work proposes an intelligent system to identify incipient faults in the electric generators of wind turbines to improve maintenance routines. Four feature extraction methods were applied to vibration signals, and different classifiers were used to predict the running status of the wind turbine. We correctly identified 94.44% of normal conditions, reducing the false positive and negative rates to 0.4% and 1.84%, respectively; a better result than other approaches already reported in the literature.

In this paper, we propose a novel approach for mobile robot localization from images. The proposal is based on supervised learning using topological representations for the environment. The whole system comprises feature extraction and classification methods. With respect to feature extraction, we consider standard methods in digital image processing, e.g. Scale-Invariant Feature Transform and Local Binary Patterns. For classification, we apply machine learning methods with rejection option. A thorough assessment of the proposal is carried out using data from virtual and real indoor environments. Additionally, we compare the proposed architectures with classic localization systems using an omnidirectional camera. Based on the results, Spatial Moments combined with Bayes classifier is the best performing model, providing high accuracy rate (99.94%) and small computational time (47.3μ s and 0.165 s for classification and extraction, respectively). Finally, we observe that localization with rejection option increases efficiency and reliability of navigation in mobile robotics.

Novel method for localization via classification with reject option using omnidirectional images.Evaluation of feature extraction and machine learning techniques in omnidirectional images.Autonomous system based on mobile robot topological map localization.Validation in real and virtual environment.Novel virtual simulation environment and two data sets (virtual and real data). Mobile robot localization, which allows a robot to identify its position, is one of main challenges in the field of Robotics. In this work, we provide an evaluation of consolidated feature extractions and machine learning techniques from omnidirectional images focusing on topological map and localization tasks. The main contributions of this work are a novel method for localization via classification with reject option using omnidirectional images, as well as two novel omnidirectional image data sets. The localization system was analyzed in both virtual and real environments. Based on the experiments performed, the Minimal Learning Machine with Nearest Neighbors classifier and Local Binary Patterns feature extraction proved to be the best combination for mobile robot localization with accuracy of 96.7% and an Fscore of 96.6%.

In later years, research in mobile robotic areas have been experiencing a growth in interest due to its vast application area. In an unknown environment, the robot's location and movement are essential for its operation. In addition, machine learning techniques, along with signal or image processing, have been applied to map the environment, locate and move the mobile robot. This article proposes a low cost and efficient approach for mobile robot localization. It uses a omnidirectional sonar with machine learning and image processing. The feature extractors used in this paper were: Structural Co-occurrence Matrix (SCM), Statistical Moments, Central Moments, Hu Moments and Gray Level Co-occurrence Matrix (GLCM). The classifiers used in this study were: Bayes classifier, k-Nearest Neighbors (kNN), Multilayer Perceptron (MLP), Optimum Path Forest (OPF) and Support Vector Machines (SVM). The results showed that the best accuracy was achieved with Central Moments as feature extractor and OPF as classifier, achieving 96.61% and with a test time of 100us.

Chronic kidney diseases cause over a million deaths worldwide every year. One of the techniques used to diagnose the diseases is renal scintigraphy. However, the way that is processed can vary depending on hospitals and doctors, compromising the reproducibility of the method. In this context, we propose an approach to process the exam using computer vision and machine learning to classify the stage of chronic kidney disease. An analysis of different features extraction methods, such as Gray-Level Co-occurrence Matrix, Structural Co-occurrence Matrix, Local Binary Patters (LBP), Hu's Moments and Zernike's Moments in combination with machine learning methods, such as Bayes, Multi-layer Perceptron, k-Nearest Neighbors, Random Forest and Support Vector Machines (SVM), was performed. The best result was obtained by combining LBP feature extractor with SVM classifier. This combination achieved accuracy of 92.00% and F1-score of 91.00%, indicating that the proposed method is adequate to classify chronic kidney disease in two stages, being a high risk of developing end-stage renal failure and other outcomes, and otherwise.

Parkinson’s disease affects millions of people around the world and consequently various approaches have emerged to help diagnose this disease, among which we can highlight handwriting exams. Extracting features from handwriting exams is an important contribution of the computational field for the diagnosis of this disease. In this paper, we propose an approach that measures the similarity between the exam template and the handwritten trace of the patient following the exam template. This similarity was measured using the Structural Cooccurrence Matrix to calculate how close the handwritten trace of the patient is to the exam template. The proposed approach was evaluated using various exam templates and the handwritten traces of the patient. Each of these variations was used together with the Naïve Bayes, OPF, and SVM classifiers. In conclusion the proposed approach was proven to be better than the existing methods found in the literature and is therefore a promising tool for the diagnosis of Parkinson’s disease.

Investigations of genetic kinship and human identification through DNA analysis of human skeletal remains have been required for several types of cases. A molecular study of this kind of sample is a challenge because of the small amount of cells available. In Brazil, there is a dominating humid tropical climate which can exert a direct influence on the microscopic morphology of bones and consequently on the DNA. The objective of this study was to analyze the microstructure of femoral compact bone tissue samples that were exposed to harsh environmental conditions in Brazil, correlating it with the amount of human DNA extracted. Compact bone fragments were used from femoral diaphysis of 20 skeletonized corpses found in the period 1998–2009 in Ribeirao Preto, Sao Paulo, Brazil. Samples were processed, stained with H&E and from the densest cellular area, 10 consecutive fields were selected by a pathologist and software Image Tool (UTSCH-USA) and Image J (NHI-USA) were used for cellular analysis. DNA extraction was performed through commercial kit and quantification with Quantifiler Duo DNA kit (Applied Biosystems). Osteocytes were observed in all cases ranging from 1 to 40 and DNA amount was correlated with the morphology observed. These results indicate preservation of osteocytes in bone material exposed to tropical environmental conditions, signifying the feasibility of obtaining DNA for genetic studies. A preliminary morphology analysis in skeleton samples can predict success in extracting DNA from these samples, since a significant correlation was found between these two variables.