Essential tremor (ET) is a progressive neurological disorder causing involuntary upper-limb oscillations that impair fine motor control and daily independence. Monitoring and attenuating such biomechanical vibrations represent a significant challenge in wearable device design. Using ET as a practical use case—characterized by oscillations in the 2–10 Hz range—this study presents the design and development of a low-cost, customizable embedded system that helps overcome that challenge. The hardware consists of a passive weighted bracelet fabricated via additive manufacturing, integrated with an ESP32-based inertial sensing module for continuous data acquisition. To reduce the microcontroller’s computational load, the system employs an onboard time-domain version of a traditional frequency-domain period extraction algorithm. The sensing architecture was evaluated through a controlled damped-harmonic-oscillator bench test, using a high-precision laboratory rotary motion sensor as the ground-truth reference. When testing the empty bracelet, the embedded accelerometer recorded a frequency of oscillation of 2.99 Hz (SD = 0.044) and decreased to 2.50 Hz (SD = 0.004) with the addition of a 137 g weight to the bracelet. The change in the oscillatory frequency shows the principle of passive inertial damping and the success in using the embedded hardware for ET monitoring.

Deep learning models require large and well-prepared datasets to achieve reliable performance. However, publicly available cardiac magnetic resonance (CMR) datasets often lack the necessary uniformity and quality for direct use in training. This paper proposes a comprehensive image preprocessing methodology aimed at improving the performance of deep learning models, specifically a 2D U-Net, for myocardium segmentation. The proposed preprocessing pipeline focuses on two main objectives: (i) enhancing image quality through contrast and brightness adjustment, gradient equalization, anisotropic diffusion filtering, and CLAHE; and (ii) accurately identifying the region of interest (ROI) containing the myocardium using a Hough Transform–based approach. After ROI detection, images are cropped to reduce dimensionality while preserving relevant anatomical structures. The methodology was evaluated using three public CMR datasets, comparing segmentation performance with and without preprocessing. The U-Net model was trained under consistent conditions, and performance was assessed using DICE and Hausdorff metrics. The preprocessing pipeline produced more uniform, higher-quality images and achieved robust ROI localization, with a 100

Eye diseases pose a significant global health challenge, affecting millions of people and leading to substantial visual impairment. Analyzing eyelid blinking and movement is crucial for monitoring patients with abnormal eyelid motions. This paper introduces the so-called Bapp (Blink Application), a mobile application designed to evaluate eyelid movement by recording the opening and closing of the eyes over time and detecting eye blinks using a machine-learning approach. The application processes pre-recorded videos and then validates the results against publicly available datasets. The proposed system operates in a unified stage, utilizing the pre-recorded video as input and evaluating the degree of eyelid openness in each frame using predictions from Google ML Kit, a machine learning-based framework integrated into the Flutter platform. The results are stored in a local database and can be exported to Excel for further analysis. Developed on the Flutter platform, the Bapp supports multiple languages, ensuring accessibility to a broad audience. The blink prediction results align with those obtained from other methods applied to the same dataset, demonstrating the app´s effectiveness in objectively monitoring the clinical progression of patients with eyelid diseases.

Blinking is a vital physiological process that protects and maintains the health of the ocular surface. Objective assessment of eyelid movements remains challenging due to the complexity, cost, and limited clinical applicability of existing tools. This study presents the Bapp (Blink Application), a mobile application developed using the Flutter framework and integrated with Google ML Kit for on-device, real-time analysis of eyelid movements, and its clinical validation. The validation was performed using 45 videos from patients, whose blinks were manually annotated by an ophthalmology specialist as the ground truth. The Bapp's performance was evaluated using standard metrics, with results demonstrating 98.4

Purpose The primary objective of this study is to identify the technology, methods, and tools commonly employed in automated applications for assessing altered eyelid movements. Methods This review consulted three databases known for their excellent reputation and technical quality: PubMed, Scopus, and Google Scholar. The initial search identified 905 papers. Removing duplicates and applying exclusion criteria narrowed the selection to 31 documents for analysis. The search included publications from January 2018 to June 2024 to focus on recent works. Results This review identified the dominance of image analysis in extracting eyelid-related parameters. The image used to analyze eyelid movements appears in 20 of the 31 selected studies. After image acquisition, machine learning is the predominant technique for extracting eyelid parameters. Recently, mobile applications have facilitated the study of eyelid movements, enabling single-step analyses and broadening accessibility. Blink frequency remains the most utilized parameter for eyelid movement analysis, and cameras are the most common sensor for capturing these movements. Conclusion The analysis identified the most used technology, methods, and tools for analyzing automated applications that assess altered eyelid movements. Specifically, this review shows that image analysis, primarily through machine learning, is vital for evaluating eyelid movements, with blink frequency as the most analyzed parameter. Also, mobile applications have expanded access to these assessments, using cameras as the primary sensor and enabling simpler, single-step analyses.

In cardiology, congenital diseases are the pathologies that most require pediatric surgical intervention. In Brazil, it is estimated that 23,000 surgeries occur per year for the treatment of congenital heart disease (CHD). These surgical procedures are complex and demand extensive training. This study presents a systematic literature review (SLR) on cardiac biomodels in surgical planning and training, identifying strategies for 3D printing and bioprinting in cardiac surgery. This SLR considered 54 studies published between 2014 and 2023. The criteria for selecting and evaluating the studies considered factors such as methods, conceptual models, and applications of biomodels produced by printing and bioprinting in surgical planning and training. Three research questions are proposed: (1) Current clinical applications of cardiac biomodels manufactured by 3D printing or bioprinting. (2) Additive manufacturing processes and materials for cardiac biomodels, including advantages and disadvantages. (3) Use of infant cardiac biomodels manufactured by 3D printing or bioprinting with application in teaching, training, and surgical planning. The evaluated studies suggest promising 3D printing or bioprinting or applications in surgical planning and training. Cardiac biomodels preemptively avoid invasive procedures. The cardiac biomodels available on the market proved to be very similar to a heart, providing surgeons with greater confidence in performing surgical procedures. Cardiac biomodels in medical universities facilitate practical classes. This review underscores the transformative potential of 3D printing and bioprinting in cardiac biomodel applications, promising enhanced surgical precision, reduced invasiveness, and an invaluable educational impact within the intricate landscape of congenital heart disease interventions.

Artificial intelligence (AI) can improve the quality control in the automotive industry. Reliable inspection methods are essential in manufacturing environments, as a single defect can damage a company’s reputation. This work presents a practical solution using computer vision and deep learning to detect small screws installed in the wheel boxes of vehicles on a moving assembly line. The system combines a custom image acquisition setup with preprocessing techniques to improve contrast and clarity, followed by object detection using the YOLOv8 algorithm. Synthetic data generated to train the model effectively allowed for more robust performance without relying solely on real-world samples. The system tested outcomes with a screw detection accuracy of 92

This systematic review aims to evaluate the applications and impacts of voice assistants (VAs) in supporting academic services. Searches explored the use of VAs in educational environments and for providing academic services in scientific databases (articles published from 2018 to 2023). The analysis of 25 articles reveals a predominance of exploratory applications and a scarcity of VA integrations with current academic systems. Concerns about data privacy and security were recurring themes, while the positive pedagogical impacts of audiovisual applications in academic services still lack solid evidence.

Learning programming logic remains an obstacle for students from different academic fields. Considered one of the essential disciplines in the field of Science and Technology, it is vital to investigate the new tools or techniques used in the teaching and learning of Programming Language. This work presents a systematic literature review (SLR) on approaches using Mobile Learning methodology and the process of learning programming in introductory courses, including mobile applications and their evaluation and validation. We consulted three digital libraries, considering articles published from 2011 to 2022 related to Mobile Learning and Programming Learning. As a result, we found twelve mobile tools for learning or teaching programming logic. Most are free and used in universities. In addition, these tools positively affect the learning process, engagement, motivation, and retention, providing a better understanding, and improving content transmission.

Os desafios no processo de aprendizagem de crianças e jovens autistas têm incentivado a busca por práticas educacionais gamificadas para esses indivíduos. Assim, o objetivo dessa Revisão Sistemática da Literatura foi investigar o papel da gamificação na aprendizagem e no estímulo da capacidade cognitiva de estudantes com Transtorno do Espectro Autista (TEA). Para isso, analisamos diversas abordagens na literatura acerca de práticas de jogos no contexto educacional com foco no autismo. Vários estudos relatam o uso da gamificação em favor da aprendizagem de estudantes com TEA.

Image denoising in medical image processing is essential to improve the visual quality of the image and, consequently, the diagnosis. Speckle, a multiplicative noise that degrades the image, is common in ultrasound imaging as it tends to degrade the resolution and contrast in the image. Several filters have been developed to reduce this kind of noise. This paper presents a comparative evaluation of some anisotropic filters used to reduce the speckle noise in ultrasound images. The evaluated filters were Coherence-Enhancing Diffusion (CED), Speckle Reducing Anisotropic Diffusion (SRAD), Detail-Preserving Anisotropic Diffusion (DPAD), and Anisotropic Diffusion Filter with Memory Based on Speckle (ADMSS). The filters processed four phantom ultrasound images with speckle noise. The comparison performances of filters occurred by using Mean Square Error, Signal to Noise Ratio, Peak Signal Noise Ratio, and Structure Similarity Index. The results showed that ADMSS better reduces speckle noises in ultrasound images, compared to the other filters evaluated.

The advance of the three-dimensional (3D) computational simulation resulted in high-fidelity virtual heart models, which are useful for implementation in medical educational tools. These computational tools help the study of cardiac anatomy and function and are of interest to many medical professionals. As the appearance of the heart surface is relevant for the comprehension of cardiac movements, the aim of the present work was the development of an animation of the external atrial and ventricular walls, during the heartbeat in a virtual 3D heart, according to physiological data. The animation used data extracted from echocardiogram images of healthy human hearts and ECG timing. The results show that displacements of the epicardial walls measured manually and wall displacements generated by the animation were comparable. Thus, the implemented animation of the atrial and ventricular external walls of the 3D virtual heart complied with those of echocardiogram images, being useful for future implementation in cardiac teaching and (or) learning tools.

Alternating current biosusceptometry assembled with anisotropic magnetoresistive sensor (AMR-ACB) is a radiation-free technique that can enable the real-time monitoring of magnetic nanoparticles in biomedical applications. The aim of this study is to assess the capability of a multi-channel AMR-ACB system in detecting maghemite nanoparticles (MN) in concentrations used in therapeutic and diagnostic (theranostic) applications, such as oncology, multimodal anticancer therapy, or magnetic resonance imaging. The axial sensibility of the AMR-ACB system was successfully characterized in a bench-top study using MN in different concentrations and distances. The MN sample was aligned with the detection axis of the AMR-ACB system, and then it was moved at fixed distances from 0 to 16 mm. The test was repeated with different MN concentrations. The results show that the AMR-ACB system is capable of detecting samples of MN with a concentration of 0.3 mg/mL, and the output signal is directly proportional to the MN concentration. The output signal exhibits exponential decaying with distance, where the signal amplitude at 8.5 mm is approximately 10% of the signal at 0 mm. This work shows that the AMR-ACB system has adequate sensibility to be employed in in vivo studies to detect MN in theranostic concentrations as well as demonstrates the potential of this magnetic method as a possible tool for future diagnosis.

The process of obtaining cardiological parameters in echocardiography images demands profound experience of the professional who analyses the images. The segmentation of the heart facilitates the obtaining of parameters in those images and benefits the diagnosis process. The present work objective was to develop a method for the segmentation of the left ventricle (LV) in echocardiography images of parasternal long-axis view from distinct databases that exhibits diversified quality. The database as a whole used in this paper consists of 67 two-dimensional gray-level echocardiograms recordings. Kohonen´s Self-Organizing Map and the Polynomial Interpolation were used to find the endocardial and epicardial walls of LV in those images. The Dice coefficient was used to compare the segmented region to the region defined by the contours drawn by a cardiologist. The mean Dice coefficient was 0.93 ± 0.02. The result validation by the Dice coefficient suggests that the system developed for the segmentation of the LV may be usable in distinct echocardiography image databases.

High-throughput medical imaging procedures, such as echocardiogram, motivate the development of new computational tools to support specialists in decision making. This paper introduces a method to detect the posterior wall of the left ventricle from echocardiogram images, considering the PLAX view. Using a Light Gradient Boosting Machine, we evaluated a classification problem and compared results with Classification and Regression Trees using k-fold cross-validation regarding five quality measures. Experimental analysis shows that the proposed approach achieves relevant median scores of 61% for Dice, 67% for Sensitivity and 98% for Specificity on a dataset of only 69 images from three different sources. With some post-processing, the median Dice score increased to 78%. Although further improvements must be considered before deploying such a tool for clinical utilization, the results presented in this paper show a promising path towards the automation of echocardiogram analysis and segmentation of the posterior wall of the left ventricle.

The myocardial infarction, known as heart attack, is the ultimate result of a prolonged/untreated cardiac ischemia. The accurate segmentation of the myocardial infarction or ischemia in images obtained from diversified sources, such as Magnetic Resonance Images or Echocardiograph, is worthwhile for the medical area or the animal experimentation. An alternative image source for ischemia/infarction segmentation is the photo, which can depict the actual heart image. This work presents a method for ischemia segmentation in rat heart photos. The method applicability was tested in pictures of human hearts available in public databases from the Internet. At first, heart images were separated from the background using GrabCut method. Secondly, the segmentation of the cardiac ischemia region was performed by using Fuzzy Clustering method. Finally, a sequence of image processing (including morphological operations to remove small components and to fill the holes) was performed to obtain the final segmentation image. All resulting images were compared with the corresponding images containing contours of cardiac ischemia drawn manually by specialists. The mean accuracy was 83.24% ± 04.16%. As for the intrinsically human errors (tracing error between two specialists: 18.94% ± 05.30%), the average accuracy is within the inter-operator variability. As for the human heart pictures obtained from public libraries, the algorithm segmented the infarction areas correctly. The results show that the algorithm effectively helps the visualization of the cardiac ischemia/infarction region and has the potential to be applied to heart images of animals or humans, representing a versatile tool to assist advances in cardiomyopathology studies.

Computational tools can be developed to help the interpretation of the ECG and the beating heart. In some cases, such tools are implemented in mobile devices and applied to “mobile learning” (m-learning) activities. In the present work, we show a mobile device tool that emulates the beating heart according to ECG data. The tool has been implemented using skeletal subspace deformation to control the movement of a 3D heart model in real time. It is possible to manipulate the heart model and visualize its inner anatomy. The results of the simulation show that the tool allows online adjustments on the virtual heart according to actual ECG data set by the user or by ECG data files (normal rhythm or arrhythmias). The comparison between clinical data and the results of the simulation validated the cardiac beating simulation. The user experience was evaluated by students of biomedical engineering using a Likert scale. The students evaluated their experience with the tool positively. The results, as a whole, show that the tool may represent a newly accessible resource to help support m-learning in the context of interpreting ECG data. Also, the tool seems to be applicable in a teaching/learning environment where undergraduates are studying either the ECG or the cardiac anatomy or function.

Echocardiographic exams allow the observation and extraction of measures related to cardiac structures. In the longitudinal parasternal view, these measures include the left ventricle end-diastolic and end-systolic diameters, end-diastolic interventricular septum thickness (IVSd), and end-diastolic left ventricle posterior wall thickness (LVPWd). Among these measures, the IVSd is important for diagnosing pathologies like hypertrophic cardiomyopathy, aneurysms, abnormal movement and structural faults. This work presents a hybrid neural network system to segment interventricular septum in echocardiographic images of parasternal longitudinal view. The hybrid system developed here consist of a Self-Organizing Map and a Multilayer Perceptron (MLP) neural network. The approach has two phases: clustering and classification. First, the Self-Organizing Map clusters image patches that are previously labeled as Septum and Non-septum. Later, an MLP is trained with information generated by the map. The MLP is then employed to classify patches of a new image resulting in a mask that indicates the probable septum regions. To validate the results, we did a semi-automatic extraction of septum thickness. The average error between the septum thicknesses obtained by the algorithm and the one manually traced was 0.5477mm ± 0.5277mm. Future recommendations are presented to improve the hybrid system performance to get more accurate results.

Cardiac rhythm disorders may cause severe heart diseases, stroke, and even sudden cardiac death. Some arrhythmias are so serious that can cause injury to other organs, for instance, brain, kidneys, lungs or liver. Therefore, early and correct diagnosis of cardiac arrhythmia is essential to the prevention of serious problems. There are expert systems to classify arrhythmias from electrocardiograms signals. However, it has been shown that not only selecting the correct features from the dataset but also generating combined features could be the key to having real progress in classification. Therefore, this paper investigates a novel hybrid evolutionary technique to perform both tasks at the same time, finding complementary features that cover different characteristics of the data. The new features were tested with a widely-used classifier called Random Forests. The method reduced a dataset with 279 attributes to 26 attributes and achieved accuracies of 86.39% for binary classification and 77.69% for multiclass. Our approach outperformed several popular feature selection, feature generation, and state-of-the-art related work from the literature.

Image segmentation of left ventricle using long-axis view of the echocardiogram is important to assist the operator in the extraction of functional parameters. The correct obtaining of this parameter can help an early diagnosis of such disease and is welcome to the medical community. However, it is not such an easy task due to the inherent equipment operator bias and the inter- and intra-observer variability. To aid in such issue, in this paper we present an automatic segmentation of the left ventricle posterior wall in echocardiographic images. Our approach employs the Self-Organizing Map to cluster the image's pixels and some image processing methods to perform the final segmentation and calculation of the left ventricle thickness. Results show that our approach, besides fully automatic, is more accurate than similar result from the literature obtained with semi-automatic methods.