Melanoma has become more widespread over the past 30 years and early detection is a major factor in reducing mortality rates associated with this type of skin cancer. Therefore, having access to an automatic, reliable system that able detect presence melanoma via dermatoscopic image lesions and/or pigmentation can be very useful tool area medical diagnosis.Among state-of-the-art methods used for automated or computer assisted diagnosis, attention should drawn Deep Learning based on...

In this work, authors address workload computation combining human activity recognition and heart rate measurements to establish a scalable framework for health at work fitness-related applications. The proposed architecture consists of two wearable sensors: one motion, another rate. system employs machine learning algorithms determine the performed by user, takes concept from ergonomics, Frimat's score, compute corresponding physical measured values providing in addition qualitative...

The dynamic features of microgrid operation, such as on-grid/off-grid operation mode, the intermittency distributed generators, and its topology due to ability reconfigure itself, cause misfiring conventional protection schemes. To solve this issue, adaptive schemes that use robust communication systems have been proposed for microgrids. However, cost solution is significantly high. This paper presented an intelligent fault detection (FD) system microgrids on basis local measurements machine...

Abstract A survey of papers in the ASEE Multidisciplinary Engineering Division for last three years shows main areas emphasis: individual courses; profiles specific projects; and capstone design courses. However, multidisciplinary education across all disciplines requires a larger-scale model that can be incorporated into any discipline, is both cost effective scalable, one fully engages benefits faculty. consortium 19 US 5 international institutions has come together around such model,...

Current global conditions and challenges in industrial manufacturing, marked by dynamism, competition, the need for responsible resource management, have increased demand sustainable manufacturing practices. The integration of Industry 4.0 recent development 5.0 added which has generated profound implications quality control process monitoring, focusing mainly on recognising patterns within environment. This study introduces a novel methodology evaluating performance pattern classification...

A system for the automatic classification of cardiac sounds can be great help doctors in diagnosis diseases. Generally speaking, main stages such systems are (i) pre-processing heart sound signal, (ii) segmentation cycles, (iii) feature extraction and (iv) classification. In this paper, we propose methods each these stages. The modified empirical wavelet transform (EWT) normalized Shannon average energy used to identify systolic diastolic intervals a recording; then, six power...

Smart networks such as active distribution network (ADN) and microgrid (MG) play an important role in power system operation. The design implementation of appropriate protection systems for MG ADN must be addressed, which imposes new technical challenges. This paper presents the validation aspects adaptive fault detection strategy based on neural (NNs) multiple sampling points MG. solution is implemented edge device. NNs are used to derive a data-driven model that uses only local...

Currently, there are many works in the literature focused on analysis of heart sounds, specifically development intelligent systems for classification normal and abnormal sounds. However, available sound databases not yet large enough to train generalized machine learning models. Therefore, is interest algorithms capable generating sounds that could augment current databases. In this article, we propose a model based generative adversary networks (GANs) generate synthetic Additionally,...

This paper deals with the problem of determining a useful energy budget for mobile robot in given environment without having to carry out experimental measures every possible exploration task. The proposed solution uses machine learning models trained on subset tasks but able make predictions untested scenarios. Additionally, model does not use any kinematic or dynamic robot, which are always available. method is based neural network hyperparameter optimization improve performance. Tabu List...

The work presented here shows a comparison between voice conversion system based on converting only the vocal tract representation of source speaker and an augmented that adds algorithm for estimating target excitation signal. estimation uses stochastic model relating signal to features. two systems were subjected objective subjective tests assessing effectiveness perceived identity overall quality synthesized speech. Male-to-male female-to- female cases tested. main this is improve...

This article presents a method that uses Linear Prediction Coefficients (LPC) and Mel-Frequency Cepstral (MFCC) as features to classify normal abnormal cardiac sounds. Three different feature vectors were tested: LPC-only, MFCC-only LPC + MFCC. Different experiments made with three classifiers: Support Vector Machine (SVM), K-Nearest Neighbor (KNN) Random Forests, using 238 samples (150 133 abnormal). Results show the best performance was obtained for combination of MFCC vectors, plus SVM...

The work presented here proposes an approach to automate the process of detecting and evaluating martial arts forms. This document presents a detection system on set static poses, which will be expanded in future into full-movement results show that it is feasible build Kinect-based stance recognition using simple machine learning algorithms, serve as basis for more complex future.

Our aim is to contribute the classification of anomalous patterns in biosignals using this novel approach. We specifically focus on melanoma and heart murmurs. use a comparative study two convolution networks Complex Real numerical domains. The idea obtain powerful approach for building portable systems early disease detection. Two similar algorithmic structures were chosen so that there no bias determined by number parameters train. Three clinical data sets, ISIC2017, PH2, Pascal, used...

This work analyzes hemodynamic phenomena within the aorta of two elderly patients and their impact on blood flow behavior, particularly affected by an endovascular prosthesis in one them (Patient II). Computational Fluid Dynamics (CFD) was utilized for this study, involving measurements velocity, pressure, wall shear stress (WSS) at various time points during third cardiac cycle, specific positions cross sections thoracic aorta. The first cross-section (Cross-Section 1, CS1) is located...

The work presented here arises from the need to find a tool for computing performance metric of given exploration algorithm under arbitrary test conditions. Such would allow most suitable predefined scenario without performing time consuming tests or simulations. is first approximation develop proposed tool, in order validate suitability our approach we choose simple and well-known algorithms scenarios, but future plan expand state-of-the-art complex scenarios. A Machine Learning system...

This paper presents a voice conversion algorithm based on Hidden Markov Models that does not requires explicit phonetic labeling of the input speech. Additionally, proposed also uses an excitation estimation previously presented by authors to achieve higher speech quality without compromising speaker identity conversion. The performance was compared, using listening tests, with recent HMM but requiring labeling. found equivalent scores while improving perceived converted Thus, as viable...

This Project explores the implementation of a computational model based on machine learning for generation synthetic melodies and harmonies, that might work as base composition musical pieces in rhythms belong to tradition culture Colombian Caribbean Region. For selection state-of-the-art analysis was performed. The implemented corresponds stack RNN layers, working upon raw audio signal. will first attempt treatment music from Region through AI, its results show it is possible synthetize...

Occupational hygiene requires evaluation of different risk sources in the workplace. The level physical workload may create stress, fatigue and injuries. Therefore, activity monitoring provides valuable information for companies assessing solving possible hazards article presents a system using wearable technology to monitor evaluate with situ measurements. uses smartwatch mobile application Android phones. During monitoring, displays physiologic variables such as heart rate, calories, body...

This work presents the results of a research project aimed at studying coding audio signals in order to obtain compact and efficient digital representation for transmission or storage purposes. A psychoacoustic model based on sub-band is implemented MATLAB®, which identifies type audio: voice music signal, human auditory system. Our focus was measuring relationship between reduced transmission/storage bitrate quality.

This paper describes a software tool for building and displaying tri-dimensional cardiac structures from sequence of Computer Tomography (CT) scan images. The development process followed this work can be divided into 5 stages. During the first stage, suitable DICOM images are selected CT machine. In second pre-processed in order to reduce noise levels enhance relevant features. third stage compares performance three image segmentation methods on images: Region growing, Otsu's algorithm...