Search

BACKGROUND: Acute ischemic stroke (AIS) is a major cause of morbidity and mortality, with hemorrhagic transformation (HT) further worsening outcomes. Traditional scoring systems have limited predictive accuracy for HT in AIS. Recent research has explored machine learning (ML) and deep learning (DL) algorithms for stroke management. This study evaluates and compares the effectiveness of ML and DL algorithms in predicting HT post-AIS, benchmarking them against conventional models. METHODS: A systematic search was conducted across PubMed, Embase, Web of Science, Scopus, and IEEE, initially yielding 1421 studies. After screening, 24 studies met the inclusion criteria. The Prediction Model Risk of Bias Assessment Tool (PROBAST) was used to assess the quality of these studies, and a qualitative synthesis was performed due to heterogeneity in the study design. RESULTS: The included studies featured diverse ML and DL algorithms, with Logistic Regression (LR), Support Vector Machine (SVM), and Random Forest (RF) being the most common. Gradient boosting (GB) showed superior performance. Median Area Under the Curve (AUC) values were 0.91 for GB, 0.83 for RF, 0.77 for LR, and 0.76 for SVM. Neural networks had a median AUC of 0.81 and convolutional neural networks (CNNs) had a median AUC of 0.91. ML techniques outperformed conventional models, particularly those integrating clinical and imaging data. CONCLUSIONS: ML and DL models significantly surpass traditional scoring systems in predicting HT. These advanced models enhance clinical decision-making and improve patient outcomes. Future research should address data expansion, imaging protocol standardization, and model transparency to enhance stroke outcomes further.

Breast cancer (BC) is a global problem, largely due to a shortage of knowledge and early detection. The speed-up process of detection and classification is crucial for effective cancer treatment. Medical image analysis methods and computer-aided diagnosis can enhance this process, providing training and assistance to less experienced clinicians. Deep Learning (DL) models play a great role in accurately detecting and classifying cancer in the huge dataset, especially when dealing with large medical images. This paper presents a novel hybrid model of DL models combined a Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) for binary breast cancer classification on two datasets available at the Kaggle repository. CNNs extract mammographic features, including spatial hierarchies and malignancy patterns, whereas LSTM networks characterize sequential dependencies and temporal interactions. Our method combines these structures to improve classification accuracy and resilience. We compared the proposed model with other DL models, such as CNN, LSTM, Gated Recurrent Units (GRUs), VGG-16, and RESNET-50. The CNN-LSTM model achieved superior performance with accuracies of 99.17% and 99.90% on the respective datasets. This paper uses prediction evaluation metrics such as accuracy, sensitivity, specificity, F-score, and the AUC curve. The results showed that our model CNN-LSTM can enhance the performance of breast cancer classifiers compared with others with 99.90% accuracy on the second dataset. [ABSTRACT FROM AUTHOR]

Artificial intelligence (AI) and deep learning (DL) techniques are increasingly used in education because of advancements in online learning platforms and their ongoing implementation. The existing methods suffer from low-processing efficiency, high prediction error, and increased memory requirements when faced with vast learning and student behavior data. Thus, based on DL, this research suggests a way to analyze student behavior in e-learning. Data on student behavior are gathered, and a learning behavior model for online learning is created. The proposed optimal DL approach aims to screen the collected behavior data using data preparation, analysis, and statistics. Additionally, the Pearson correlation coefficient (PCC) approach is employed to determine the degree of data similarity. The novelty of the research is followed by utilizing an optimized DL network, known as a deep neural network with red deer optimization (ODNN-RDO), to mine students’ behavior data in e-learning programs. Two datasets, metrics including accuracy, precision, and recall, together with error measures like relative error, the root mean square error (RMSE), and absolute error, are utilized to test the created models. The improved generated models achieved 98.15% accuracy and 0–0.04% error compared to the current methods. The optimization procedure subsequently optimizes the components to acquire the best outcomes regarding faculty and parent performance monitoring of students. With effective monitoring, this model maximizes the e-learning platform for planning student growth. [ABSTRACT FROM AUTHOR]

Background and Objectives: Computer-aided diagnostic systems have achieved remarkable success in the medical field, particularly in diagnosing malignant tumors, and have done so at a rapid pace. However, the generalizability of the results remains a challenge for researchers and decreases the credibility of these models, which represents a point of criticism by physicians and specialists, especially given the sensitivity of the field. This study proposes a novel model based on deep learning to enhance lung cancer diagnosis quality, understandability, and generalizability. Methods: The proposed approach uses five computed tomography (CT) datasets to assess diversity and heterogeneity. Moreover, the mixup augmentation technique was adopted to facilitate the reliance on salient characteristics by combining features and CT scan labels from datasets to reduce their biases and subjectivity, thus improving the model's generalization ability and enhancing its robustness. Curriculum learning was used to train the model, starting with simple sets to learn complicated ones quickly. Results: The proposed approach achieved promising results, with an accuracy of 99.38%; precision, specificity, and area under the curve (AUC) of 100%; sensitivity of 98.76%; and F1-score of 99.37%. Additionally, it scored a 00% false positive rate and only a 1.23% false negative rate. An external dataset was used to further validate the proposed method's effectiveness. The proposed approach achieved optimal results of 100% in all metrics, with 00% false positive and false negative rates. Finally, explainable artificial intelligence (XAI) using Gradient-weighted Class Activation Mapping (Grad-CAM) was employed to better understand the model. Conclusions: This research proposes a robust and interpretable model for lung cancer diagnostics with improved generalizability and validity. Incorporating mixup and curriculum training supported by several datasets underlines its promise for employment as a diagnostic device in the medical industry. [ABSTRACT FROM AUTHOR]

The Healthcare Internet of Things (H-IoT) is a rapidly developing problem solving model with significant potential to improve patient care and healthcare outcomes. This study focuses on integrating cryptographic platforms into H-IoT systems to enable secure data access. We present insights on how to manage challenges such as cyber risks, resource constraints, latency, and energy consumption by exploring cryptographic approaches in diverse H-IoT applications. We explore important topics including big data management, blockchain, machine learning, edge computing, and software-defined networks. Additionally, we examine real-time operational challenges and emerging trends, such as remote patient monitoring and predictive analytics. Our research underscores the need for strong encryption mechanisms, access controls, device authentication, and proactive threat detection to improve the safety and efficiency of healthcare services. By combining cryptographic approaches with the architecture of the H-IoT system, this work provides the foundation for resilient healthcare infrastructures. Addressing current challenges and anticipating future opportunities contribute to strategic healthcare planning that prioritizes patient privacy and data security while anticipating future strategic healthcare planning opportunities. This article aims to provide valuable information to researchers by covering energy-efficient and resource-optimized healthcare system approaches that are integrated into the development of H-IoT systems. [ABSTRACT FROM AUTHOR]

Background: Acute ischemic stroke (AIS) is a major cause of morbidity and mortality, with hemorrhagic transformation (HT) further worsening outcomes. Traditional scoring systems have limited predictive accuracy for HT in AIS. Recent research has explored machine learning (ML) and deep learning (DL) algorithms for stroke management. This study evaluates and compares the effectiveness of ML and DL algorithms in predicting HT post-AIS, benchmarking them against conventional models. Methods: A systematic search was conducted across PubMed, Embase, Web of Science, Scopus, and IEEE, initially yielding 1421 studies. After screening, 24 studies met the inclusion criteria. The Prediction Model Risk of Bias Assessment Tool (PROBAST) was used to assess the quality of these studies, and a qualitative synthesis was performed due to heterogeneity in the study design. Results: The included studies featured diverse ML and DL algorithms, with Logistic Regression (LR), Support Vector Machine (SVM), and Random Forest (RF) being the most common. Gradient boosting (GB) showed superior performance. Median Area Under the Curve (AUC) values were 0.91 for GB, 0.83 for RF, 0.77 for LR, and 0.76 for SVM. Neural networks had a median AUC of 0.81 and convolutional neural networks (CNNs) had a median AUC of 0.91. ML techniques outperformed conventional models, particularly those integrating clinical and imaging data. Conclusions: ML and DL models significantly surpass traditional scoring systems in predicting HT. These advanced models enhance clinical decision-making and improve patient outcomes. Future research should address data expansion, imaging protocol standardization, and model transparency to enhance stroke outcomes further. [ABSTRACT FROM AUTHOR]

The using of Electroencephalography (EEG) signals for motor imagery (MI) has recently gained significant attention due to their remarkable ability to detect an individual's intention to perform specific actions. MI signals have proven useful in enabling individuals with disabilities to control devices such as wheelchairs through neural commands, and have even expanded into applications like autonomous driving. Therefore, ensuring accurate classification of MI tasks from EEG signals is crucial for the development of a reliable Brain-Computer Interface (BCI) system. This article introduces a novel approach to classifying MI tasks using Deep Learning (DL) techniques. The proposed methodology encompasses several steps, including data preprocessing, feature extraction using Common Spatial Pattern (CSP) and Wavelet Packet Decomposition (WPD), and the evaluation of four distinct classifiers. These classifiers involve combinations of two, three, four, and five Convolutional Neural Networks (CNNs). Empirical evaluations highlight the effectiveness of employing five CNNs, which yield the most favorable results. Our approach demonstrates promising performance metrics such as accuracy, precision, recall, and F1 score. Specifically, the method achieves accuracy, precision, recall, and F1 score values of 64.75%, 64.94%, 65.63%, and 64.13%, respectively, indicating its potential to enhance the accuracy of MI task classification within EEG signals. [ABSTRACT FROM AUTHOR]

Copyright of Retos: Nuevas Perspectivas de Educación Física, Deporte y Recreación is the property of Federacion Espanola de Asociaciones de Docentes de Educacion Fisica and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

In the dynamic field of ophthalmology, artificial intelligence (AI) is emerging as a transformative tool in managing complex conditions like uveitis. Characterized by diverse inflammatory responses, uveitis presents significant diagnostic and therapeutic challenges. This systematic review explores the role of AI in advancing diagnostic precision, optimizing therapeutic approaches, and improving patient outcomes in uveitis care. A comprehensive search of PubMed, Scopus, Google Scholar, Web of Science, and Embase identified over 10,000 articles using primary and secondary keywords related to AI and uveitis. Rigorous screening based on predefined criteria reduced the pool to 52 high-quality studies, categorized into six themes: diagnostic support algorithms, screening algorithms, standardization of Uveitis Nomenclature (SUN), AI applications in management, systemic implications of AI, and limitations with future directions. AI technologies, including machine learning (ML) and deep learning (DL), demonstrated proficiency in anterior chamber inflammation detection, vitreous haze grading, and screening for conditions like ocular toxoplasmosis. Despite these advancements, challenges such as dataset quality, algorithmic transparency, and ethical concerns persist. Future research should focus on developing robust, multimodal AI systems and fostering collaboration among academia and industry to ensure equitable, ethical, and effective AI applications. The integration of AI heralds a new era in uveitis management, emphasizing precision medicine and enhanced care delivery. [ABSTRACT FROM AUTHOR]

The modern digital world is experiencing growing security risks, with cyber threats happening more often and becoming more complicated. These threats affect both businesses and individuals. Machine learning (ML) and deep learning (DL) have emerged as vital tools in cybersecurity, enabling the analysis of extensive datasets to identify potential cyber threats. This study proposes a novel technique utilizing ML and DL algorithms for threat detection. We began with data preprocessing, which included cleansing the data, addressing missing values through Multiple Imputation by Chained Equations (MICE), and applying transformations such as encoding and standard scaling. To address class imbalance, we employed the Synthetic Minority Over-sampling Technique (SMOTE). For feature selection, we used Forward Feature Elimination (FFE), Backward Feature Elimination (BFE), and Recursive Feature Elimination (RFE) to identify the most relevant features. We trained three ML classifiers: Support Vector Machine (SVM), Naïve Bayes (NB), and K-Nearest Neighbors (KNN), along with three DL models: Long Short-Term Memory (LSTM), Recurrent Neural Network (RNN), and Convolutional Neural Network (CNN). Model performance was evaluated using metrics such as accuracy, precision, recall, and F1 score, alongside loss graphs and confusion matrices. The highest accuracy at 99% was attained by the LSTM model, while the CNN demonstrated superior precision (98%), recall (97%), and F1 score (97%). Additionally, the CNN, RNN, and SVM models achieved an accuracy of 98%. These results illustrate the effectiveness of ML and DL in identifying cybersecurity threats, highlighting their potential to enhance defenses against emerging cyber risks.

Coronary computed tomography angiography (CCTA) is a noninvasive imaging modality of cardiac structures and vasculature considered comparable to invasive coronary angiography for the evaluation of coronary artery disease (CAD) in several major cardiovascular guidelines. Conventional image acquisition, processing, and analysis of CCTA imaging have progressed significantly in the past decade through advances in technology, computation, and engineering. However, the advent of artificial intelligence (AI)-driven analysis of CCTA further drives past the limitations of conventional CCTA, allowing for greater achievements in speed, consistency, accuracy, and safety. AI-driven CCTA (AI-CCTA) has achieved a significant reduction in radiation exposure for patients, allowing for high-quality scans with sub-millisievert radiation doses. AI-CCTA has demonstrated comparable accuracy and consistency in manual coronary artery calcium scoring against expert human readers. An advantage over invasive coronary angiography, which provides luminal information only, CCTA allows for plaque characterization, providing detailed information on the quality of plaque and offering further prognosticative value for the management of CAD. Combined with AI, many recent studies demonstrate the efficacy, accuracy, efficiency, and precision of AI-driven analysis of CCTA imaging for the evaluation of CAD, including assessing degree stenosis, adverse plaque characteristics, and CT fractional flow reserve. The limitations of AI-CCTA include its early phase in investigation, the need for further improvements in AI modeling, possible medicolegal implications, and the need for further large-scale validation studies. Despite these limitations, AI-CCTA represents an important opportunity for improving cardiovascular care in an increasingly advanced and data-driven world of modern medicine.

With the growing complexity of skin disorders and the challenges of traditional diagnostic methods, AI offers exciting new solutions that can enhance the accuracy and efficiency of dermatological assessments. Reflectance Confocal Microscopy (RCM) stands out as a non-invasive imaging technique that delivers detailed views of the skin at the cellular level, proving its immense value in dermatology. The manual analysis of RCM images, however, tends to be slow and inconsistent. By combining artificial intelligence (AI) with RCM, this approach introduces a transformative shift toward precise, data-driven dermatopathology, supporting more accurate patient stratification, tailored treatments, and enhanced dermatological care. Advancements in AI are set to revolutionize this process. This paper explores how AI, particularly Convolutional Neural Networks (CNNs), can enhance RCM image analysis, emphasizing machine learning (ML) and deep learning (DL) methods that improve diagnostic accuracy and efficiency. The discussion highlights AI’s role in identifying and classifying skin conditions, offering benefits such as a greater consistency and a reduced strain on healthcare professionals. Furthermore, the paper explores AI integration into dermatological practices, addressing current challenges and future possibilities. The synergy between AI and RCM holds the potential to significantly advance skin disease diagnosis, ultimately leading to better therapeutic personalization and comprehensive dermatological care.

In the realm of image-based breast cancer detection and severity assessment, this study delves into the revolutionary potential of sophisticated artificial intelligence (AI) techniques. By investigating image processing, machine learning (ML), and deep learning (DL), the research illuminates their combined impact on transforming breast cancer diagnosis. This integration offers insights into early identification and precise characterization of cancers. With a foundation in 125 research articles, this article presents a comprehensive overview of the current state of image-based breast cancer detection. Synthesizing the transformative role of AI, including image processing, ML, and DL, the review explores how these technologies collectively reshape the landscape of breast cancer diagnosis and severity assessment. An essential aspect highlighted is the synergy between advanced image processing methods and ML algorithms. This combination facilitates the automated examination of medical images, which is crucial for detecting minute anomalies indicative of breast cancer. The utilization of complex neural networks for feature extraction and pattern recognition in DL models further enhances diagnostic precision. Beyond diagnostic improvements, the abstract underscores the substantial influence of AI-driven methods on breast cancer treatment. The integration of AI not only increases diagnostic precision but also opens avenues for individualized treatment planning, marking a paradigm shift toward personalized medicine in breast cancer care. However, challenges persist, with issues related to data quality and interpretability requiring continued research efforts. Looking forward, the abstract envisions future directions for breast cancer identification and diagnosis, emphasizing the adoption of explainable AI techniques and global collaboration for data sharing. These initiatives promise to propel the field into a new era characterized by enhanced efficiency and precision in breast cancer care.

Harini Sriraman, Saleena Badarudeen, Saransh Vats, Prakash Balasubramanian School of Computer Science and Engineering, Vellore Institute of Technology, Chennai, 600127, IndiaCorrespondence: Prakash Balasubramanian, School of Computer Science Engineering, Vellore Institute of Technology, Chennai, Tamil Nadu, India, Tel +91-044-39931228, Fax +91-044-39932555, Email Prakash.bala@vit.ac.inAbstract: Deep Learning (DL) drives academics to create models for cancer diagnosis using medical image processing because of its innate ability to recognize difficult-to-detect patterns in complex, noisy, and massive data. The use of deep learning algorithms for real-time cancer diagnosis is explored in depth in this work. Real-time medical diagnosis determines the illness or condition that accounts for a patient’s symptoms and outward physical manifestations within a predetermined time frame. With a waiting period of anywhere between 5 days and 30 days, there are currently several ways, including screening tests, biopsies, and other prospective methods, that can assist in discovering a problem, particularly cancer. This article conducts a thorough literature review to understand how DL affects the length of this waiting period. In addition, the accuracy and turnaround time of different imaging modalities is evaluated with DL-based cancer diagnosis. Convolutional neural networks are critical for real-time cancer diagnosis, with models achieving up to 99.3% accuracy. The effectiveness and cost of the infrastructure required for real-time image-based medical diagnostics are evaluated. According to the report, generalization problems, data variability, and explainable DL are some of the most significant barriers to using DL in clinical trials. Making DL applicable for cancer diagnosis will be made possible by explainable DL.Keywords: artificial intelligence, AI, machine learning, DL, CNN, healthcare, real-time diagnosis, classification, image processing, elastography, feedforward neural network

Synthetic Aperture Radar (SAR) imaging plays a vital role in maritime surveillance, yet the detection of small vessels poses a significant challenge when employing conventional Constant False Alarm Rate (CFAR) techniques, primarily due to the limitations in resolution and the presence of clutter. Deep learning (DL) offers a promising alternative, yet it still struggles with identifying small targets in complex SAR backgrounds because of feature ambiguity and noise. To address these challenges, our team has developed the AFSC network, which combines anti-aliasing techniques with fully shared convolutional layers to improve the detection of small targets in SAR imagery. The network is composed of three key components: the Backbone Feature Extraction Module (BFEM) for initial feature extraction, the Neck Feature Fusion Module (NFFM) for consolidating features, and the Head Detection Module (HDM) for final object detection. The BFEM serves as the principal feature extraction technique, with a primary emphasis on extracting features of small targets, The NFFM integrates an anti-aliasing element and is designed to accentuate the feature details of diminutive objects throughout the fusion procedure, HDM is the detection head module and adopts a new fully shared convolution strategy to make the model more lightweight. Our approach has shown better performance in terms of speed and accuracy for detecting small targets in SAR imagery, surpassing other leading methods on the SSDD dataset. It attained a mean Average Precision ( A P ) of 69.3% and a specific A P for small targets ( A P S ) of 66.5%. Furthermore, the network's robustness was confirmed using the HRSID dataset. [ABSTRACT FROM AUTHOR]

ABSTRACT The detection of brain tumours presents a significant challenge in the medical domain, where prompt and precise diagnosis is crucial as patient outcomes depend on it. Conventional deep neural networks perform well in carrying out various imaging tasks within the healthcare sector; however, their effectiveness often falls short of expectations in practical applications due to the substantial computational resources required and issues with reliability. In this research, an optimised and effective deep learning model founded on the DenseNet‐169 architecture is introduced for the classification of magnetic resonance imaging brain tumours, which is particularly advantageous for smart healthcare systems and information and communication technology (ICT) settings with limited computational capabilities. The model compression methodologies, including pruning and quantization, have been employed to significantly diminish the dimensions and intricacy of the model while achieving a classification accuracy of 97.07%. Furthermore, this endeavour necessitates the enhancement of the model's interpretability through the utilisation of explainable artificial intelligence methodologies such as Gradient‐weighted Class Activation Mapping (Grad‐CAM) and SHapley Additive exPlanations (SHAP), which will aid clinicians in highlighting crucial areas of the images and validating feature importance concerning the decisions rendered by the model. A comparative performance evaluation is conducted against DenseNet‐169, ResNet‐50 and various other models to delineate the superior efficacy of our model, rendering it exceptionally adept for knowledge‐driven, real‐time brain tumour diagnosis within smart healthcare and ICT systems where resources are constrained. [ABSTRACT FROM AUTHOR]

Successful medical treatment for patients with COVID-19 requires rapid and accurate diagnosis. Fighting the COVID-19 pandemic requires an automated system to diagnose the virus on Chest X-Ray (CXR) images. CXR images are frequently used in healthcare as they offer the potential for rapid and accurate disease diagnosis. SARS-CoV-2 targets the respiratory system, resulting in pneumonia with additional symptoms, such as dry cough, fatigue, and fever, which could be misdiagnosed as pneumonia, TB, or lung cancer. There is difficulty in differentiating the features of COVID-19 from other diseases that have similarities in CXR images. Automated Computer-Aided Diagnosis (CAD) systems incorporate machine or deep learning methods to improve efficiency and accuracy. CNNs are among the most widely used methods, as they have shown encouraging accuracy in identifying COVID-19 in CXR images. This study presents a hybrid deep learning model to provide faster diagnosis of COVID-19 infection using CXR images. The Densenet201 model was used for feature extraction and a Multi-Layer Perceptron (MLP) was used for classification. The proposed method achieved 98.82 % accuracy and similar sensitivity, specificity, precision, recall, and F1 score. These results are promising when compared to other DL models trained in similar datasets. [ABSTRACT FROM AUTHOR]

This research presents an innovative approach to disease diagnosis from medical imaging data, specifically focusing on X-ray images of osteoporosis. The proposed method revolves around a Transformer-based Guided Convolutional Neural Network (TGCNN), which integrates the spatial awareness of CNNs with the sophisticated relationship modeling capabilities of Transformers. The model is meticulously trained, achieving a training accuracy of 96% and a testing accuracy of 92%, with corresponding training and testing losses of 0.0034 and 0.0364, respectively. To gauge its effectiveness, TGCNN is rigorously compared with established models, namely ResNet50, AlexNet, and DenseNet169, showcasing superior performance. ResNet50 achieves a training accuracy of 86% and testing accuracy of 82%, with losses of 0.064 and 0.174. AlexNet demonstrates a training accuracy of 83% and testing accuracy of 78%, with losses of 0.04 and 0.64. DenseNet169 attains a training accuracy of 92%, testing accuracy of 88%, and losses of 0.0674 and 0.094. The comprehensive results and comparative analysis affirm the efficacy of the proposed TGCNN model for accurate and efficient disease diagnosis from X-ray images. This work sets the stage for further advancements in medical imaging diagnostics, with potential applications in diverse datasets and real-time diagnostic scenarios. Future work will explore optimizations and extensions to enhance the model's versatility and performance across various medical imaging domains. [ABSTRACT FROM AUTHOR]

Streamflow prediction offers crucial information for managing water resources, flood control, and hydropower generation. Yet, reliable streamflow prediction is challenging due to the complexity and nonlinearity of the rainfall-runoff relationship. This study investigated the comparative performance of the newly integrated self-attention-based deep learning (DL) model, SA-Conv1D-BiGRU with Conv1D-LSTM, and bidirectional long short-term memory (Bi-LSTM) models for streamflow prediction under different time-series conditions, and a range of variable input combinations based on flood events. All datasets passed quality control procedures, and the time lag for generating input series was established through Pearson correlation analysis. 80% of the data was used for training, whereas 20% was used to evaluate the model's performance. The performance of the models was evaluated using three metrics: mean absolute error (MAE), root mean square error (RMSE), and correlation coefficient (R²). The findings reveal the excellent potential of DL models for streamflow prediction, with the SA-Conv1D-BiGRU model outperforming other models under different time-series characteristics. Despite the complexity, the Conv1D-LSTM models did not outperform the Bi-LSTM model. In conclusion, the results are condensed into themes of model variability and time-series characteristics. Consequently, different architectures in DL models had a greater influence on streamflow prediction accuracy than input time lags and time-series features. [ABSTRACT FROM AUTHOR]

Huanglongbing (HLB; citrus greening) and citrus canker are invasive diseases afflicting Florida citrus groves, causing financial losses through yield loss, smaller fruit, blemishes, premature fruit drop and/or eventual tree death. Often, symptoms of these diseases resemble those of other disease or disorders. Early detection of HLB and canker via in-grove leaf inspection can permit more effective mitigation tactics and management of groves. Autonomous, vision-based disease scouting in a grove offers a financial benefit to the Florida citrus industry. This study investigates the potential of hyperspectral reflectance imagery (HSI) for detecting and classifying these conditions in the presence of other, less consequential leaf defects. Leaves with visible symptoms of HLB, canker, zinc deficiency, scab, melanose, greasy spot, and a control set were collected and both sides were imaged with a line-scan hyperspectral camera. Spectral bands from this imagery were selected using two methods: an unsupervised method based on principal component analysis (PCA), a supervised method based on linear discriminant analysis (LDA), which were compared with randomly selected bands as a control. The YOLOv8 network architecture was trained to classify each side of these leaves with each band combination. LDA-selected bands from the front of the leaves yielded an overall classification accuracy of 87.09 %, with higher recall of melanose and precision of control than any other model tested. Leaves with HLB and zinc deficiency were classified most accurately, with both band selection methods yielding F1 scores of at least 0.955 and 0.934, respectively. These findings favor the use of supervised band selection for HSI-based in-grove disease detection.

ObjectiveTo develop automated algorithms for the detection of posterior vitreous detachment (PVD) using OCT imaging.DesignEvaluation of a diagnostic test or technology.SubjectsOverall, 42 385 consecutive OCT images (865 volumetric OCT scans) obtained with Heidelberg Spectralis from 865 eyes from 464 patients at an academic retina clinic between October 2020 and December 2021 were retrospectively reviewed.MethodsWe developed a customized computer vision algorithm based on image filtering and edge detection to detect the posterior vitreous cortex for the determination of PVD status. A second deep learning (DL) image classification model based on convolutional neural networks and ResNet-50 architecture was also trained to identify PVD status from OCT images. The training dataset consisted of 674 OCT volume scans (33 026 OCT images), while the validation testing set consisted of 73 OCT volume scans (3577 OCT images). Overall, 118 OCT volume scans (5782 OCT images) were used as a separate external testing dataset.Main outcome measuresAccuracy, sensitivity, specificity, F1-scores, and area under the receiver operator characteristic curves (AUROCs) were measured to assess the performance of the automated algorithms.ResultsBoth the customized computer vision algorithm and DL model results were largely in agreement with the PVD status labeled by trained graders. The DL approach achieved an accuracy of 90.7% and an F1-score of 0.932 with a sensitivity of 100% and a specificity of 74.5% for PVD detection from an OCT volume scan. The AUROC was 89% at the image level and 96% at the volume level for the DL model. The customized computer vision algorithm attained an accuracy of 89.5% and an F1-score of 0.912 with a sensitivity of 91.9% and a specificity of 86.1% on the same task.ConclusionsBoth the computer vision algorithm and the DL model applied on OCT imaging enabled reliable detection of PVD status, demonstrating the potential for OCT-based automated PVD status classification to assist with vitreoretinal surgical planning.Financial disclosuresProprietary or commercial disclosure may be found after the references.

PurposeTo compare the diagnostic accuracy and explainability of a Vision Transformer deep learning technique, Data-efficient image Transformer (DeiT), and ResNet-50, trained on fundus photographs from the Ocular Hypertension Treatment Study (OHTS) to detect primary open-angle glaucoma (POAG) and identify the salient areas of the photographs most important for each model's decision-making process.DesignEvaluation of a diagnostic technology.Subjects participants and controlsOverall 66 715 photographs from 1636 OHTS participants and an additional 5 external datasets of 16 137 photographs of healthy and glaucoma eyes.MethodsData-efficient image Transformer models were trained to detect 5 ground-truth OHTS POAG classifications: OHTS end point committee POAG determinations because of disc changes (model 1), visual field (VF) changes (model 2), or either disc or VF changes (model 3) and Reading Center determinations based on disc (model 4) and VFs (model 5). The best-performing DeiT models were compared with ResNet-50 models on OHTS and 5 external datasets.Main outcome measuresDiagnostic performance was compared using areas under the receiver operating characteristic curve (AUROC) and sensitivities at fixed specificities. The explainability of the DeiT and ResNet-50 models was compared by evaluating the attention maps derived directly from DeiT to 3 gradient-weighted class activation map strategies.ResultsCompared with our best-performing ResNet-50 models, the DeiT models demonstrated similar performance on the OHTS test sets for all 5 ground-truth POAG labels; AUROC ranged from 0.82 (model 5) to 0.91 (model 1). Data-efficient image Transformer AUROC was consistently higher than ResNet-50 on the 5 external datasets. For example, AUROC for the main OHTS end point (model 3) was between 0.08 and 0.20 higher in the DeiT than ResNet-50 models. The saliency maps from the DeiT highlight localized areas of the neuroretinal rim, suggesting important rim features for classification. The same maps in the ResNet-50 models show a more diffuse, generalized distribution around the optic disc.ConclusionsVision Transformers have the potential to improve generalizability and explainability in deep learning models, detecting eye disease and possibly other medical conditions that rely on imaging for clinical diagnosis and management.

Content Based Image Retrieval (CBIR) plays a significant role in identifying the similarity of images with large datasets. It is identified based on the size, colour, and texture features of the image. But in such conditions, it is complex to determine the features of query images in large datasets and does not show accurate similarity when compared with every image in the retrieval process. In order to perform an efficient similarity of images, a novel Machine Learning (ML) approach Kernelized Radial Basis Auto-Encoder Function Neural Network (Ker_RadBAEFNN) technique is proposed that performs the individual image classification in the retrieval process. Moreover, the neural networks are optimised based on the reinforcement process and perform the extraction process regarding individual images. Further,reinforcement-based optimisation estimates the images in neural networks for undertaking an automatic feature extraction of query images. The performance of the classification process is validated based on MNIST, METU, and COCO datasets that determined the efficiency of the recognition and classification process of image retrieval. The experimental analysis is carried out based on various measures such as accuracy, precision, recall, F1-score, RMSE, and MAPE for the proposed and existing GLCM-ABC, PSO-ANN, IRB-CNN, FAGWO, and OCAM methods. The analysis shows that the performance of the proposed attained better effectiveness with attained accuracy by 98% and diminished for state-of-the-art techniques as 92%, 95%, 94%, 96.8%, as well as 96%, respectively. Compared to existing methods, the accuracy rate of the proposed method is maximised by 1.3%. [ABSTRACT FROM AUTHOR]

The field of wireless communication systems has experienced significant advancements in recent years, leading to the emergence of two promising technologies: non-orthogonal multiple access (NOMA) and deep learning (DL)-based autoencoders (AE). Through power allocation, NOMA enables multiple users to share a single frequency band, while AE can compress and decompress data with high precision. Integrating NOMA and AE enables end-to-end (E2E) transmission with a superior signal-to-noise (SNR) ratio. To further enhance the wireless network's block error rate (BLER) performance, the multiple-input, multiple-output (MIMO) technique is also incorporated into the newly proposed system. With the incorporation of the MIMO signal, the system is abbreviated as the MIMO-NOMA-AE system. The suggested technique for detecting MIMO-NOMA-AE signals has demonstrated a remarkable performance gain in SNR surpassing the traditional successive interference cancellation (SIC)-based NOMA system. The proposed system also performs better than previously utilized deep neural network (DNN)-based SISO-NOMA-AE communication systems. While this method shows potential for future wireless communication systems, further research and testing are necessary to evaluate its practical feasibility and effectiveness. [ABSTRACT FROM AUTHOR]

This research mainly explores the existing drowning detection methodologies, focusing primarily on the roles carried out by Machine Learning (ML) and Deep Learning (DL) algorithms. It directly emphasizes the dominance of ML in the analysis of raw sensor data along with the contribution of DL to computer vision, which also reveals the present gap between advanced vision along detection models. The holistic approaches are mainly advocated, potentially integrating wearable devices, vision-based systems, as well as sensors while also balancing their performance, regional applicability, and cost-effectiveness. The challenges aligned to enabling real-time detection and reduced latency are important for the time-sensitive realm of incidents related to drowning. Future directions necessarily include the exploration of advanced forms of vision models and segmentation techniques for innovative detection algorithms. Integration of wearable devices and sensors with the inclusion of vision-based systems is important for the required adaptability. The upcoming proposal aims to integrate robotics into rescue operations bringing revolution to response times. The study also covers the requirement for a compact combination of ML and DL algorithms and a generalized solution for the equilibrium maintenance between cost-effectiveness, sophistication, and regional applicability. [ABSTRACT FROM AUTHOR]

Email communication has become integral to various industries, but the pervasive issue of spam emails poses significant challenges for service providers. This research proposes a study leveraging Machine Learning (ML) and Deep Learning (DL) techniques to effectively classify spam emails. Methods such as Logistic Regression (LR), Naïve Bayes (NB), Random Forest (RF), and Artificial Neural Networks (ANNs) are employed to construct robust models for accurate spam detection. By amalgamating these techniques, the aim is to enhance efficiency and precision in spam detection, aiding email and IoT service providers in mitigating the detrimental effects of spam. Evaluation of the proposed models revealed promising outcomes. LR, RF, and NB achieved an impressive accuracy of 97% and an F1-Score of 97.5%, showcasing their efficacy in accurately identifying spam emails. The ANN model demonstrated slightly superior performance, with 98% accuracy and 97.5% F1-score, suggesting potential improvements in accuracy and robustness in spam filtering systems. These findings underscore the viability of both traditional ML algorithms and DL approaches in addressing the challenges of email spam classification, paving the way for more effective spam detection mechanisms in electronic communication platforms. [ABSTRACT FROM AUTHOR]

The paper proposes a novel approach called MusicEmo, a transformer-based intelligent system for music emotion generation and recognition. The paper highlights the challenges of creating emotionally resonant music that is musically cohesive and diverse. The proposed approach addresses this challenge by introducing a theme-based conditioning approach, which trains the transformer to manifest the conditioning sequence as thematic material that appears multiple times in the generated result. The MusicEmo architecture incorporates an emotion vector and an LSTM model for creating symbolic musical sequences that are musically coherent and emotionally resonant. The proposed framework outperforms state-of-the-art approaches based on musical consistency and emotional resonance. The transformer-based approach offers a fresh and original way of creating music based on emotions, and it can potentially revolutionize how we create and experience music in the future. [ABSTRACT FROM AUTHOR]

In line with recent advances in neural drug design and sensitivity prediction, we introduce a novel architecture for the interpretable prediction of anticancer compound sensitivity utilizing a multimodal attention-based convolutional encoder. Our approach is based on three primary foundations: prior knowledge of intracellular interactions from protein-protein interaction networks, gene expression profiles of tumors, and the structure of chemicals as a SMILES sequence. With R2 = 0.86 and RMSE = 0.89, our multi-scale convolutional attention-based encoder significantly outperforms a baseline model trained on Morgan fingerprints, a set of SMILES-based encoders, and the previously reported state-of-the-art for multimodal drug sensitivity prediction. Talk about the Ensemble Convolution Neural Network Model: A Novel Regression-Based Approach (ECNN-NRNN) to Drug Sensitivity Analysis Using Multiple Pharmaomics Data Sets and Addressing Heterogeneity in Feature Selection for Sub-Pharmacoomics Parameters. Because some pharmacogenomic data is available online and should be made publicly available, it is essential to address drug sensitivity prediction and drug identification and design. Outline how the performance in sensitivity prediction can be improved using conventional methods, and provide an experimental evaluation. [ABSTRACT FROM AUTHOR]

Copyright of Paradigma is the property of Universidad Pedagogica Experimental Libertador and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Efficient techniques for classifying diabetes types are essential for effective treatment of this chronic condition. Artificial intelligence (AI) algorithms, particularly Support Vector Machine (SVM) algorithms, have shown great promise in this area. A recent study utilized SVMs with different functions, including linear, polynomial, and sigmoid, to categorize diabetes into distinct types. The dataset was divided into training and test sets to perform prediction analysis, with the SVM classifier applied to both sets. The proposed system excels in segregating data for processing, training the model to accurately differentiate between Type 1 and Type 2 diabetes. By leveraging AI-based algorithms, the system enhances the precision of diabetes classification, thereby facilitating more tailored and effective treatment plans for patients. The implementation of SVM algorithms in diabetes classification not only demonstrates the capabilities of AI in medical diagnostics but also underscores the potential for integrating advanced computational techniques in healthcare. The study's approach ensures that the classifier can handle diverse patient data, improving its generalizability and robustness. As diabetes cases continue to rise globally, such AI-driven solutions could play a pivotal role in early diagnosis, personalized treatment, and ongoing management, ultimately contributing to better health outcomes and reduced healthcare costs. [ABSTRACT FROM AUTHOR]

This study investigates the effectiveness of the DistilBERT model in classifying tweets related to disasters. This study achieved significant predictive accuracy through a comprehensive analysis of the dataset and iterative refinement of the model, including adjustments to hyperparameters. The benchmark model developed highlights the benefits of DistilBERT, with its reduced size and improved processing speed contributing to greater computational efficiency while maintaining over 95% of BERT's capabilities. The results indicate an impressive average training accuracy of 92.42% and a validation accuracy of 82.11%, demonstrating the practical advantages of DistilBERT in emergency management and disaster response. These findings underscore the potential of advanced transformer models to analyze social media data, contributing to better public safety and emergency preparedness. [ABSTRACT FROM AUTHOR]