AI Sens., Volume 1, Issue 1 (September 2025) – 6 articles

Age-related macular degeneration (AMD) is the leading cause of blindness in developed countries. Predicting its progression is crucial for preventing late-stage AMD, as it is an irreversible retinal disease. Both genetic factors and retinal images are instrumental in diagnosing and predicting AMD progression. Previous studies have explored automated diagnosis using single fundus images and genetic variants, but they often fail to utilize the valuable longitudinal data from multiple visits. Longitudinal retinal images offer a dynamic view of disease progression, yet standard Long Short-Term Memory (LSTM) models assume consistent time intervals between training and testing, limiting their effectiveness in real-world settings. To address this limitation, we propose time-varied Long Short-Term Memory (TV-LSTM), which accommodates irregular time intervals in longitudinal data. Our innovative approach enables the integration of both longitudinal fundus images and AMD-associated genetic variants for more precise progression prediction. Our TV-LSTM model achieved an AUC-ROC of 0.9479 and an AUC-PR of 0.8591 for predicting late AMD within two years, using data from four visits with varying time intervals. Full article

Optical sensors have undergone significant evolution, transitioning from discrete optical microsystems toward sophisticated photonic integrated circuits (PICs) that leverage artificial intelligence (AI) for enhanced functionality. This review systematically explores the integration of optical sensing technologies with AI, charting the advancement from conventional optical microsystems to AI-driven smart devices. First, we examine classical optical sensing methodologies, including refractive index sensing, surface-enhanced infrared absorption (SEIRA), surface-enhanced Raman spectroscopy (SERS), surface plasmon-enhanced chiral spectroscopy, and surface-enhanced fluorescence (SEF) spectroscopy, highlighting their principles, capabilities, and limitations. Subsequently, we analyze the architecture of PIC-based sensing platforms, emphasizing their miniaturization, scalability, and real-time detection performance. This review then introduces the emerging paradigm of in-sensor computing, where AI algorithms are integrated directly within photonic devices, enabling real-time data processing, decision making, and enhanced system autonomy. Finally, we offer a comprehensive outlook on current technological challenges and future research directions, addressing integration complexity, material compatibility, and data processing bottlenecks. This review provides timely insights into the transformative potential of AI-enhanced PIC sensors, setting the stage for future innovations in autonomous, intelligent sensing applications. Full article

This study presents CORE-ReID V2, an enhanced framework built upon CORE-ReID V1. The new framework extends its predecessor by addressing unsupervised domain adaptation (UDA) challenges in person ReID and vehicle ReID, with further applicability to object ReID. During pre-training, CycleGAN is employed to synthesize diverse data, bridging image characteristic gaps across different domains. In the fine-tuning, an advanced ensemble fusion mechanism, consisting of the Efficient Channel Attention Block (ECAB) and the Simplified Efficient Channel Attention Block (SECAB), enhances both local and global feature representations while reducing ambiguity in pseudo-labels for target samples. Experimental results on widely used UDA person ReID and vehicle ReID datasets demonstrate that the proposed framework outperforms state-of-the-art methods, achieving top performance in mean average precision (mAP) and Rank-k Accuracy (Top-1, Top-5, Top-10). Moreover, the framework supports lightweight backbones such as ResNet18 and ResNet34, ensuring both scalability and efficiency. Our work not only pushes the boundaries of UDA-based object ReID but also provides a solid foundation for further research and advancements in this domain. Full article

Deficits in social interaction and communication characterize Autism Spectrum Disorder (ASD). Although widely recognized by its symptoms, diagnosing ASD remains challenging due to its wide range of clinical presentations. Methods: In this study, we propose a method to assist in the early diagnosis of autism, which is currently primarily based on clinical assessments. Our approach aims to develop an early differential diagnosis based on electroencephalogram (EEG) signals, seeking to identify patterns associated with ASD. In this study, we used EEG data from 56 participants obtained from the Sheffield dataset, including 28 individuals diagnosed with Autism Spectrum Conditions (ASC) and 28 neurotypical controls, applying numerical techniques to handle missing data. Subsequently, after a detailed analysis of the signals, we applied three different starting approaches: one with the original database and the other two with selection of the most significant attributes using the PSO and evolutionary search methods. In each of these approaches, we applied a series of machine learning models, where relatively high performances for classification were observed. Results: We achieved accuracies of 99.13% ± 0.44 for the dataset with original signals, 99.23% ± 0.38 for the dataset after applying PSO, and 93.91% ± 1.10 for the dataset after the evolutionary search methodology. These results were obtained using classical classifiers, with SVM being the most effective among the first two approaches, while Random Forest with 500 trees proved more efficient in the third approach. Conclusions: Even with all the limitations of the base, the results of the experiments demonstrated promising findings in identifying patterns associated with Autism Spectrum Disorder through the analysis of EEG signals. Finally, we emphasize that this work is the starting point for a larger project with the objective of supporting and democratizing the diagnosis of ASD both in children early and later in adults. Full article

(1) Background: The current uses of smartwatch wearable devices have expanded, not only being a part of everyday routine life but also playing a dynamic role in the early detection of many behavioral patterns of users. Furthermore, in the modern era, there is an increasing trend of mental disturbances even in early adolescence, a phenomenon that continues into academic life. Taking into account the current situation, the objective of this systematic literature review emphasizes the role of AI wearable devices in the early symptom detection of burnout in the student population. (2) Methods: A systematic literature review was designed based on the PRISMA guidelines. The general extracted aspect was to exploit all the current related research evidence about the effectiveness of wearable devices in the student population. (3) Results: The reviewed studies document the importance of physiological monitoring and AI-driven predictive models, with the collaboration of self-reported scales in assessing mental well-being. It is reported that stress is the most frequently studied burnout-related symptom. Meanwhile, heart rate (HR) and heart rate variability (HRV) are the most commonly used biomarkers that can be used to monitor and evaluate early burnout detection. (4) Conclusions: Despite the promising potential of these technologies, several challenges and limitations must be addressed to enhance their effectiveness and reliability. Full article