Background : Electroencephalography (EEG) interpretation for epilepsy diagnosis faces persistent challenges including specialist shortages, variable interpretation accuracy, and limited accessibility. Artificial intelligence (AI)-based automated interpretation systems promise to address these limitations, yet their diagnostic performance compared to clinical experts remains incompletely characterized. Objective: To systematically evaluate and meta -analyze the diagnostic accuracy of AI-enabled EEG interpretation compared with human clinical experts for epilepsy detection. Methods : We conducted a systematic review following PRISMA-DTA and Cochrane MECIR standards, searching PubMed, Scopus, IEEE Xplore, and Google Scholar through June 2025. Studies directly comparing AI algorithms with human expert interpretation using identical EEG datasets were included. Quality assessment employed QUADAS-AI criteria. Bivariate meta -analysis was performed to estimate pooled sensitivity, specificity, diagnostic odds ratios, and likelihood ratios. It is important to note that AI algorithms evaluated in these studies were trained using expert-labeled EEG data through supervised learning approaches; our comparison focuses on diagnostic accuracy outcomes rather than algorithmic independence from expert knowledge. Results : Three studies encompassing 9,775 EEG examinations met inclusion criteria. AI demonstrated superior pooled sensitivity (0.83–0.85 vs. 0.77–0.80) and specificity (0.83–0.85 vs. 0.72–0.75) compared to human experts. The diagnostic odds ratio for AI was approximately double that of humans (12–13 vs. 6–7). AI exhibited consistently narrower confidence intervals, indicating greater interpretive reliability. For normal versus abnormal EEG classification, AI achieved 86–90% sensitivity with enhanced consistency compared to human evaluators. However, substantial heterogeneity (I 2 > 75%) and methodological limitations were identified across studies. Conclusions: AI-based EEG interpretation demonstrates diagnostic performance equal or superior to human experts with enhanced consistency, supporting potential implementation as clinical triage tools. However, limited transparency, patient selection bias, and deployment feasibility constraints warrant further investigation before widespread clinical adoption. Multiple sources of uncertainty affect AI-based diagnostic systems in clinical applications. Internal uncertainties include model parameter uncertainty, threshold selection variability, and training data limitations. External uncertainties encompass population heterogeneity, EEG acquisition variability, and clinical context differences. Parametric uncertainties arise from model architecture choices, while non-parametric uncertainties reflect distribution-free variations in real-world data. The amounts and structures of these uncertainties are often unknown in operational settings, necessitating robust uncertainty quantification methods for safe clinical deployment.

This paper presents an analysis of Gaussian distribution for radar signals modeling, offering a robust framework for characterizing clutter and target echoes. For radar systems, the generalized likelihood ratio test (GLRT) detectors emerge as versatile tools for navigating diverse statistical conditions through maximum likelihood estimation (MLE). This paper extends the previously available closed-form expressions for the probability of false alarm (PFA) and probability of detection (PD) for GLRT, considering fluctuating target signals. We show that, despite variations in system models, the same closed-form expressions can be obtained. This extension reaffirms the resilience and applicability of the proposed methodology in enhancing radar signal analysis. Derived expressions are confronted with Monte Carlo simulations, confirming the tightness of the expressions obtained.

This study investigates the application of Tensor Robust Principal Components analysis (TRPCA) as a pre-processing tool in classifying oil rigs using synthetic aperture radar (SAR) images. The pre-processing considers the tensor composition of original images and subsequent attribute extraction using the VGG-16 convolutional neural network from the low-rank and sparse images. The extracted features are then classified using Support Vector Machine (SVM), Neural Network (NET), and Logistic Regression (LR) classifiers. The experimental evaluation utilized C-band VH-polarization SAR images from the Sentinel-1 system. The findings indicate that TRPCA-based pre-processing enhances classification accuracy, outperforming existing methods documented in the literature.

In the context of synthetic aperture radar (SAR) systems, the Gamma distribution stands out as a strong contender for clutter modeling. The cell-averaging constant false alarm rate (CA-CFAR) detector has frequently been employed as a well-balanced detection technique, ensuring a blend of high performance and comparatively low complexity. In this study, we evaluate, in an exact manner, the CA-CFAR detector’s performance in an SAR system, assuming the presence of correlated Gamma-distributed clutter. To this aim, we derive novel exact expressions for the probability of detection (PD) and the probability of false alarm (PFA), simplifying their runtime evaluations without depending on specific mathematical software. The independent and identically distributed (IID) and independent not identically distributed (INID) cases are also analyzed. In particular, for the IID case, our derived PD and PFA expressions are given in closed form. Our analytical findings are validated through Monte Carlo simulations.

Wavelength-resolution (WR) synthetic aperture radar (SAR) change detection (CD) has been used to detect concealed targets in forestry areas. However, most proposed methods are generally based on matrix or vector analyses and, therefore, do not exploit information embedded in multidimensional data. In this letter, a CD method for WR SAR image stacks based on tensor robust principal component analysis (TRPCA) is proposed. The proposed CD method used the new tensor nuclear norm induced by the definition of the tensor-tensor product to exploit temporal and spatial information contained in the image stack. To assess the performance of the proposed method, we considered SAR images obtained by the very high frequency (VHF) WR CARABAS-II SAR system. Experiments for three different stack sizes show that a significant performance gain can be achieved when large image stacks are considered. The proposed CD method performs better in terms of probability of detection (PD) and false alarm rate (FAR) than the other five CD methods in VHF WR SAR images, including one based on matrix robust principal component analysis (RPCA). In a particular setting, it achieves a PD of 99% and a FAR of 0.028 false alarms per km2.

We propose an algorithm based on Information Theory to detect changes in Ultra-Wideband (UWB) Very-High Frequency (VHF) Synthetic Aperture Radar (SAR) images with high performance and low complexity. Our algorithm models the clutter-plus-noise using six different distributions and computes a scalar statistic for each pixel based on a multi-temporal stack of images. With this statistic, it is then possible to apply hypothesis testing and classification methods to infer the occurrence of a change. In this context, we derive expressions necessary for the entropy-based statistics, including the entropy variance for the Weibull and Rice distributions. We also evaluate the computational time complexity of the algorithm for each distribution studied. Furthermore, a masking strategy is used to reduce false alarms significantly. We show that the mask mapping assumptions are mild in scenarios with stacks of images, allowing its use in many scenarios. Our algorithm achieves a false alarm rate (FAR) of 0.08 and a probability of detection (PD) of 100%, outperforming existing methods on the CARABAS II data set.

This letter presents two new change detection (CD) methods for synthetic aperture radar (SAR) image stacks based on the Neyman–Pearson criterion. The first proposed method uses the data from wavelength–resolution images stack to obtain background statistics, which are used in a hypothesis test to detect changes in a surveillance image. The second method considersa prioriinformation about the targets to obtain the target statistics, which are used together with the previously obtained background statistics, to perform a hypothesis test to detect changes in a surveillance image. A straightforward processing scheme is presented to test the proposed CD methods. To assess the performance of both proposed methods, we considered the coherent all radio band sensing (CARABAS)-II SAR images. In particular, to obtain the temporal background statistics required by the derived methods, we used stacks with six images. The experimental results show that the proposed techniques provide a competitive performance in terms of probability of detection and false alarm rate compared with other CD methods.

This letter presents a clutter statistical analysis for stacks of wavelength-resolution synthetic aperture radar (SAR) images. Each image stack consists of SAR images generated by the same sensor, using the same flight track illuminating the same scene but with a time separation between the illuminations. We test three candidate statistical distributions for time changes in the stack, namely, Rician, Rayleigh, and log-normal. The tests results reveal that the Rician distribution is a very good candidate for modeling stack of wavelength-resolution SAR images, where 98.59% of the tested samples passed the Anderson-Darling (AD) goodness-of-fit test. Also, it is observed that the presence of changes in the ground scene is related to the tested samples that have failed in the AD test for the Rician distribution hypothesis.

This paper presents a study on the constant false alarm rate (CFAR) filter design for change detection algorithms (CDA). More specifically, we are interested in CFAR filters used in CDA for low frequency ultra-wideband (UWB) synthetic aperture radar (SAR) systems. The filter design performance was evaluated in terms of false alarm rate (FAR) and probability of detection (PD). For evaluation purposes, we considered a set of SAR images obtained with the CARABAS-II system. The results are compared with the ones presented in [1], where the same CDA was considered, except for the CFAR filter. The results show that relevant FAR performance improvements can be obtained by just modifying the CFAR filter parameters taking into account the image resolution and target characteristics.

In this paper we propose a low-complexity codebook-based beamforming with four transmit antennas and quantized feedback channel. The codebook design aggregates the effect of power allocation and phase rotation through a simple quantized transmit scheme. The codebook-based beamforming uses the feedback information in order to maximize the instantaneous signal-to-noise ratio (SNR) at the receiver. As a result, the proposed scheme presents an array gain. An SNR analysis is performed and it is used to find the optimal feedback information in the sense of maximizing the instantaneous SNR. A bit error rate (BER) analysis for a quantized feedback channel is also derived and it is used to compare to the results obtained for the proposed scheme under different levels of quantization. Simulations are performed over quasi-static flat Rayleigh fading channels for different closed-loop codebook-based schemes with four transmit antennas and unitary transmission rate. Results illustrate that the proposed scheme achieves full diversity order and outperforms other good schemes in terms of array gain.