Search

The timely detection and accurate localization of brain tumors is crucial in preserving people’s quality of life. Thankfully, intelligent computational systems have proven invaluable in addressing these challenges. In particular, the UNET model can extract essential pixel-level features to automatically identify the tumor’s location. However, known deep learning-based works usually directly feed the 3D volume into the model, which causes excessive computational complexity. This paper presents an approach to boost the UNET network, reducing computational workload while maintaining superior efficiency in locating brain tumors. This concept could benefit portable or embedded recognition systems with limited resources for operating in real time. This enhancement involves an automatic slice selection from the MRI T2 modality volumetric images containing the most relevant tumor information and implementing an adaptive learning rate to avoid local minima. Compared with the original model (7.7 M parameters), the proposed UNET model uses only 2 M parameters and was tested on the BraTS 2017, 2020, and 2021 datasets. Notably, the BraTS2021 dataset provided outstanding binary metric results: 0.7807 for the Intersection Over the Union (IoU), 0.860 for the Dice Similarity Coefficient (DSC), 0.656 for the Sensitivity, and 0.9964 for the Specificity compared to vanilla UNET.

Advances in the field of Brain-Computer Interfaces (BCIs) aim, among other applications, to improve the movement capacities of people suffering from the loss of motor skills. The main challenge in this area is to achieve real-time and accurate bio-signal processing for pattern recognition, especially in Motor Imagery (MI). The significant interaction between brain signals and controllable machines requires instantaneous brain data decoding. In this study, an embedded BCI system based on fist MI signals is developed. It uses an Emotiv EPOC+ Brainwear®, an Altera SoCKit® development board, and a hexapod robot for testing locomotion imagery commands. The system is tested to detect the imagined movements of closing and opening the left and right hand to control the robot locomotion. Electroencephalogram (EEG) signals associated with the motion tasks are sensed on the human sensorimotor cortex. Next, the SoCKit processes the data to identify the commands allowing the controlled robot locomotion. The classification of MI-EEG signals from the F3, F4, FC5, and FC6 sensors is performed using a hybrid architecture of Convolutional Neural Networks (CNNs) and Long Short-Term Memory (LSTM) networks. This method takes advantage of the deep learning recognition model to develop a real-time embedded BCI system, where signal processing must be seamless and precise. The proposed method is evaluated using k-fold cross-validation on both created and public Scientific-Data datasets. Our dataset is comprised of 2400 trials obtained from four test subjects, lasting three seconds of closing and opening fist movement imagination. The recognition tasks reach 84.69% and 79.2% accuracy using our data and a state-of-the-art dataset, respectively. Numerical results support that the motor imagery EEG signals can be successfully applied in BCI systems to control mobile robots and related applications such as intelligent vehicles.

A hybrid Linear Quadratic Regulator (LQR) and Proportional-Integral (PI) control for a MicroGrid (MG) under unbalanced linear and nonlinear loads was presented and evaluated in this paper. The designed control strategy incorporates the microgrid behavior, low-cost LQR, and error reduction in the stationary state by the PI control, to reduce the overall energetic cost of the classical PI control applied to MGs. A Genetic Algorithm (GA) calculates the parameters of LQR with high-accuracy fitness function to obtain the optimal controller parameters as settling time and overshoot. The gain values of the classical PI controller were determined through the improved LQR values and geometrical root locus. When MG operates in the grid-tied mode under unbalanced conditions, the controller performance of the Current Source Inverter (CSI) of the MG is considerably affected. Consequently, the CSI operates in a negative-sequence mode to compensate for unbalanced current at the Point of Common Coupling (PCC) between the MG and the utility grid. The study cases involved the reduction of the negative-sequence percentage in the current at the PCC, mitigation of harmonics in the current signal injected by the MG, and close related power quality issues. All these cases have been analyzed by implementing an MG connected at the PCC of a low-voltage distribution network. A numerical model of the MG in Matlab/Simulink was implemented to verify the performance of the designed LQR-PI control to mitigate or overcome the power quality concerns. The extensive simulations have permitted verifying the robustness and effectiveness of the proposed strategy.

Early detection of different levels of tremors helps to obtain a more accurate diagnosis of Parkinson’s disease and to increase the therapy options for a better quality of life for patients. This work proposes a non-invasive strategy to measure the severity of tremors with the aim of diagnosing one of the first three levels of Parkinson’s disease by the Unified Parkinson’s Disease Rating Scale (UPDRS). A tremor being an involuntary motion that mainly appears in the hands; the dataset is acquired using a leap motion controller that measures 3D coordinates of each finger and the palmar region. Texture features are computed using sum and difference of histograms (SDH) to characterize the dataset, varying the window size; however, only the most fundamental elements are used in the classification stage. A machine learning classifier provides the final classification results of the tremor level. The effectiveness of our approach is obtained by a set of performance metrics, which are also used to show a comparison between different proposed designs.

The continuous advances in intelligent systems and cutting-edge technology have greatly influenced the development of intelligent vehicles. Recently, integrating multiple sensors in cars has improved and spread the advanced drive-assistance systems (ADAS) solutions for achieving the goal of total autonomy. Despite current self-driving approaches and systems, autonomous driving is still an open research issue that must guarantee the safety and reliability of drivers. This work employs images from two cameras and Global Positioning System (GPS) data to propose a 3D vision-based object localization and classification method for assisting a car during driving. The experimental platform is a prototype of a two-sitter electric vehicle designed and assembled for navigating the campus under controlled mobility conditions. Simultaneously, color and depth images from the primary camera are combined to extract 2D features, which are reprojected into 3D space. Road detection and depth features isolate point clouds representing the objects to construct the occupancy map of the environment. A convolutional neural network was trained to classify typical urban objects in the color images. Experimental tests validate car and object pose in the occupancy map for different scenarios, reinforcing the car position visually estimated with GPS measurements.

Short circuits occurring between turns within the windings are widely known as one of the primary causes of damage in electrical transformers; as a result, early detection plays a fundamental role in preventing further and more serious damage. This study introduces a novel approach that relies on the analysis of current and vibration signals, specifically employing the analysis of quaternion signals, to effectively detect short circuits within electrical transformers., offering an identification of conditions ranging from a healthy state to six levels of short circuit turns. in a no-load transformer, i.e., 0, 5, 10, 15, 20, 25 and 30 SCT. This proposed method employs quaternion rotation to extract statistical features that can be used to classify the condition of the transformer. To evaluate the effectiveness of the proposed methodology, an experimental validation is carried out using a 1.5 kVA transformer, comparing its performance against other existing methods. The results demonstrate the feasibility of the proposal, accurately identifying various levels of SCT, achieving an accuracy of 97.5%, using only 100 samples with the k nearest neighbors method.

The lifetime of induction motors can be significantly extended by installing diagnostic systems for monitoring their operating conditions. In particular, detecting broken bar failures in motors is important for avoiding the risk of short circuits or other accidents with serious consequences. In the literature, many approaches have been proposed for motor fault detection; however, additional generalized methods based on local and statistical analysis could provide a low-complexity and feasible solution in this field of research. The proposed work presents a methodology for detecting one or two broken rotor bars using the sums and differences histograms (SDH) and machine learning classifiers in this context. From the SDH computed in one phase of the motor’s current, nine texture features are calculated for different displacements. Then, all features are used to train two classifiers and to find the best displacements for faults and health identification in the induction motors. A final experimental evaluation considering the best displacements shows an accuracy of 98.16% for the homogeneity feature and a few signal samples used in a decision tree classifier. Additionally, a polynomial regression curve validates the use of 50 samples to obtain an accuracy of 88.15%, whereas the highest performance is achieved for 250 samples.

The light polarization properties provide relevant information about linear–optical media quality and condition. The Stokes–Mueller formalism is commonly used to represent the polarization properties of the incident light over sample tests. Currently, different Stokes Polarimeters are mainly defined by resolution, acquisition rate, and light to carry out accurate and fast measurements. This work presents the implementation of an automatic Stokes dynamic polarimeter to characterize non-biological and biological material samples. The proposed system is configured to work in the He-Ne laser beam’s reflection or transmission mode to calculate the Mueller matrix. The instrumentation stage includes two asynchronous photoelastic modulators, two nano-stepper motors, and an acquisition data card at 2% of accuracy. The Mueller matrix is numerically calculated by software using the 36 measures method without requiring image processing. Experiments show the efficiency of the proposed optical array to calculate the Mueller matrix in reflection and transmission mode for different samples. The mean squared error is calculated for each element of the obtained matrix using referenced values of the air and a mirror. A comparison with similar works in the literature validates the proposed optical array.

Category suggestions or recommendations for customers or users have become an essential feature for commerce or leisure websites. This is a growing topic that follows users’ activity in social networks generating a huge quantity of information about their interests, contacts, among many others. These data are usually collected to analyze people’s behavior, trends, and integrate a complete user profile. In this sense, we analyze a dataset collected from Pinterest to predict the gender and age by processing input images using a Convolutional Neural Network. Our method is based on the meaning of the image rather than the visual content. Additionally, we propose a heuristic-based approach for text analysis to predict users’ age and gender from Twitter. Both of the classifiers are based on text and images and they are compared with various similar approaches in the state of the art. Suggested categories are based on association rules conformed by the activity of thousands of users in order to estimate trends. Computer simulations showed that our approach can recommend interesting categories for a user analyzing his current interest and comparing this interest with similar users’ profiles or trends and, therefore, achieve an improved user profile. The proposed method is capable of predicting the user’s age with high accuracy, and at the same time, it is able to predict gender and category information from the user. The certainty that one or more suggested categories be interesting to people is higher for those users with a large number of publications.