This systematic review examines EEG-based imagined speech classification, emphasizing directional words essential for development in the brain–computer interface (BCI). This study employed a structured methodology to analyze approaches using public datasets, ensuring systematic evaluation and validation of results. This review highlights the feature extraction techniques that are pivotal to classification performance. These include deep learning, adaptive optimization, and frequency-specific decomposition, which enhance accuracy and robustness. Classification methods were explored by comparing traditional machine learning with deep learning and emphasizing the role of brain lateralization in imagined speech for effective recognition and classification. This study discusses the challenges of generalizability and scalability in imagined speech recognition, focusing on subject-independent approaches and multiclass scalability. Performance benchmarking across various datasets and methodologies revealed varied classification accuracies, reflecting the complexity and variability of EEG signals. This review concludes that challenges remain despite progress, particularly in classifying directional words. Future research directions include improved signal processing techniques, advanced neural network architectures, and more personalized, adaptive BCI systems. This review is critical for future efforts to develop practical communication tools for individuals with speech and motor impairments using EEG-based BCIs.

Predictive learning outcomes’ models for online students can provide useful information for instructors to estimate students’ final performance in the early stages of a course. Anticipating student performance can improve learning efficiency. Most previous research models analyzing student data have focused on handcrafted features. Handcraft feature-based models have limitations in exploring new behavioral patterns that indicate students’ performance and how they can be used in online courses. The clickstream data contain a significant amount of information that accurately describes students’ learning processes, which makes it difficult to construct using hand-crafted features. To analyze student behavior effectively, we attempted to capture critical knowledge from the Natural Language Processing (NLP) field to the field of student performance prediction in Massive Open Online Courses (MOOCs), owing to how closely they resemble one another. In this article, we propose a novel framework for automatically producing useful data representation that enhances prediction outcomes using student learning behavior clickstream data with a self-supervised learning approach. First, we developed a self-supervised clickstream pre-training setup to model learner click generation. Second, we adjusted these latent representations before applying them to a downstream supervised learning task. Extensive experimental results on two real-world datasets demonstrated that the proposed approach is effective. The combined approach of skip-gram embeddings with Principal Component Analysis (PCA) achieved the highest accuracy, particularly on the Xutangx dataset, with an accuracy of approximately 72.70% and an F1-score of approximately 81.03%. Furthermore, when applied to the KDDCUP dataset, this methodology exhibited an even higher performance, with an accuracy of 80.91% and an F1-score of 87.42%. Our results showed the potential of NLP techniques to improve dropout prediction in MOOCs by extracting informative representations from clickstream data, enabling a deeper understanding of student behavior, and facilitating early intervention strategies.

Classifying imagined speech from electroencephalographic signals remains a significant challenge in brain–computer interface research due to the low amplitude, nonstationary properties, and high intersubject variability of the neural activity involved. However, a key unresolved problem is the poor generalizability of subject-independent imagined-speech EEG classification, largely due to inadequate multiscale signal representations and suboptimal integration of feature extraction, selection, and classification. This study introduces a subject-independent classification framework that integrates multiscale signal decomposition, statistical feature selection, and ensemble classification. Specifically, Rational-Dilation Wavelet Transform is employed to achieve flexible time-frequency analysis, followed by the extraction of time-domain, energy-based, and fractal features from each subband. Statistically discriminative features are identified using one-way analysis of variance, and classification is performed using an Extreme Gradient Boosting model with automated hyperparameter tuning. Evaluated on the Thinking Out Loud dataset, a publicly available imagined speech corpus, the proposed pipeline achieved a subject-independent test set accuracy of 64.45%, substantially outperforming prior methods that typically reported accuracies between 25.8% and 33.9%. The cross-validation results closely matched the performance of the test set, indicating a strong generalization. These findings demonstrate the effectiveness of signal-aware decomposition and statistically principled feature engineering in improving the robustness of imagined speech classification. The proposed approach provides a promising basis for future developments in noninvasive neural decoding and silent communication BCIs.

With increasing crowd sizes nowadays, the risks of overcrowding, including injuries, accidents, and even fatalities, have become a great concern. Motivated by the need for safer public spaces, this work designed and developed Amneen, a crowd management system that uses Artificial Intelligence (AI) and Computer Vision (CV). The system provides authorities and event organizers with a real-time tool to track crowd density in public places and prevent dangerous situations before they occur. Amneen integrates two AI models: a head detection model using YOLO-11 and an overcrowding prediction model using Stochastic Gradient Descent Regressor (SGDRegressor). Using live video footage from installed cameras, the system estimates the number of people in a specific area, displays crowd statistics through an interactive dashboard, and sends early warnings when the situation worsens. Additionally, by analyzing historical data patterns, the system predicts congestion before it occurs. The detection model demonstrated strong performance in real-time, processing an image in 6.5 ms with a precision of 93.95%, a recall of 90.91%, an F1-score of 92.41%, and a mean Average Precision (mAP) of 96.26%. The prediction model yielded an MAE of 18.11 and an score of 0.53, indicating moderate predictive performance. Unit and usability testing demonstrated effectiveness and ease of use, highlighting its potential to improve the general quality of life.

Currently, rapid growth in airports is encountered. Thus, enhancing customer experience, safety, and satisfaction using advanced technologies, such as deep learning (DL), is important. This research study explores the available DL and image processing algorithms to identify the most suitable algorithms that provide high detection accuracy to support the lost and found units in airports. Using the collected data from the surveillance camera technology in the airports, the performance of the algorithms is benchmarked and evaluated on an identical dataset to acquire an understanding of the distinctive features of each algorithm, the ability to distinguish between the algorithms and choose the approach to object identification that works best in each situation. Faster R-CNN and YOLO-v8 algorithms are the focus of this study by exemplifying the various abilities of both and identifying their suitability to be used in individual and object identification in airports to support lost and found units. YOLO-v8 is a better fit for the airport lost and found units to help passengers allocate their lost bags than Faster R-CNN. This is evident from the evaluation results of measuring the accuracy and speed of both algorithms, in which YOLO-v8 performed significantly better. Considering the focus of this research and the need for real-time object detection, YOLO-v8 is a superior fit for airport use as the analysis results are real-time, which is preferable in such an industry.