Search Results (350)

Adaptive bitrate streaming over HTTP relies on carefully constructed bitrate ladders and ordered sets of bitrate–resolution pairs to deliver optimal perceptual quality under fluctuating network conditions. While content-aware methods based on convex hull optimisation have substantially improved ladder efficiency for Video-on-Demand, they require exhaustive multi-resolution pre-encoding that is computationally prohibitive under the real-time constraints of live streaming. This challenge is compounded by the H.266/Versatile Video Coding (VVC) standard, which offers approximately 50% compression gains over HEVC at 8–10× the encoding complexity. This paper presents LiveCH-VVC, a latency-aware dynamic bitrate ladder prediction framework for VVC-encoded live streaming over Low-Latency DASH (LL-DASH) with CMAF packaging. The framework introduces four integrated modules: (i) a Lightweight Dual-Path CNN (LDP-CNN), obtained via teacher–student knowledge distillation (∼5 M parameters, 148 ms GPU inference), that jointly extracts spatial–temporal features from raw frames and compression-domain statistics from a fast VVC probe encode; (ii) an adaptive scene change detector with exponential moving average thresholding (F1 = 0.925) that triggers ladder updates only upon significant complexity shifts; (iii) a temporally augmented XGBoost multi-label classifier that predicts latency-constrained Pareto-optimal bitrate–resolution pairs; and (iv) an online adaptation engine that integrates Common Media Client Data (CMCD) feedback from CDN edge servers for continuous closed-loop refinement. Comprehensive evaluation on 81 UHD sequences (∼4050 CMAF segments) from three benchmark datasets demonstrates an average BD-Rate of $+0.68\%$ relative to the per-segment oracle convex hull $5.4\times$ better than the state-of-the-art ARTEMIS framework ($+3.67\%$) while achieving 73.3% encoding time savings, 2.37 s end-to-end latency, and a QoE score of 81.6 in live simulation with 100 concurrent clients. Ablation analysis confirms that the dual-path compression-domain branch ($+0.44$ pp) and temporal context augmentation ($+0.35$ pp) are the primary performance drivers, while the online adaptation mechanism provides $42\%$ relative improvement over extended streaming sessions. Full article

Aiming at core bottlenecks in weak and small target detection in radar range-Doppler (RD) spectra under low signal-to-noise ratio (SNR)—including severe performance degradation of traditional constant false alarm rate (CFAR) detectors and the inherent trade-off difficulty faced by existing deep learning methods in balancing detection accuracy, localization precision, and real-time performance—this paper proposes a progressive sub-pixel-level intelligent detection algorithm named F²DN-CCWL. The algorithm constructs a three-stage detection pipeline: global candidate screening, local fine discrimination, and weighted localization, and implements a full-stack customized design covering network architecture, soft-label training strategy, and post-processing modules. Simulation and field-measured results demonstrate that at −20 dB SNR, the proposed algorithm achieves a detection probability of 95.3%, a false alarm rate of 3.1%, an average localization error of 0.76 pixels, and a single-frame inference latency of 47.21 ms. This method offers a high-performance engineering solution for radar-based detection of low observable targets. Full article

Accurate R-peak detection in electrocardiogram (ECG) signals is fundamental for cardiovascular analysis. However, most existing methods address differences in sampling frequency (f_s) through signal resampling or transfer learning, which may alter the temporal definition of annotated events. In this study, we propose a f_s consistent framework for ECG R-peak detection that avoids both resampling and retraining. The proposed method is based on low-sampling morphological learning combined with physiological temporal constraints (PTC). A lightweight classifier based on Extreme Gradient Boosting (XGB) was trained on 128-Hz ECG data from the MIT-BIH Normal Sinus Rhythm Database to learn local morphological structures, and feature extraction is defined in milliseconds with time-normalized derivatives to ensure consistency across f_s. The trained model is directly applied to higher-f_s datasets (360 Hz, 500 Hz, and 1000 Hz) without modification. Final peak locations are determined through deterministic processing, including PTC and local snap processing. Experimental results demonstrated that the proposed method achieved stable detection performance across multiple sampling frequencies. When evaluated in a sample-wise manner, the proposed method achieved mean F1-scores of 0.885 on MIT-BIH Arrhythmia Database (360 Hz), 0.848 on Lobachevsky University Electrocardiography Database (LUDB, 500 Hz, sinus rhythm), 0.837 on LUDB (500 Hz, arrhythmia), and 0.953 on PTB Diagnostic ECG Database (1000 Hz), without any resampling or retraining. The integration of probabilistic candidate detection and deterministic temporal alignment enables consistent peak localization under cross-frequency conditions. These findings demonstrate that augmenting machine learning with deterministic decision mechanisms provides a principled framework for f_s-consistent ECG peak detection. Full article

Photovoltaic systems have recently attracted significant attention for the free, clean, and sustainable energy they generate. In this work, thermal image processing techniques were developed and utilized to classify hot spots on solar photovoltaic panels. Thermal images were classified into three categories: (a) ideal images, where images do not contain hot spots; (b) images affected by shadow; and (c) images affected by bird drops. The proposed classification was developed using image processing techniques, including histogram analysis, contrast enhancement, and filtering tools. The attained classes are then matched to the decrease in electrical power output. The proposed method was applied to thermal images to detect and classify the target hot spot. Experimental results showed that the estimated error was approximately 6.3% of the total number of images used in the research, with error rates of 6.57% for the shadow hot spot type and 6.67% for the bird drops (mud-like class). Moreover, the accuracy of the proposed method was around 93.7%. Full article

Intelligent fault diagnosis of rotating machinery typically assumes that training and test data share the same operating speed, an assumption that rarely holds in industrial end-of-line testing, where a rotor must be certified across a range of shaft speeds. In this paper, we expose this assumption through a systematic benchmark of four deep learning architectures (TCN, 1D-CNN, BiLSTM, and CNN-BiLSTM) on a laboratory rotor testbench with three operating speeds (1000, 2000, and 3000 rpm) and four unbalance fault classes. Under within-speed 5-fold cross-validation, all four models achieve a perfect macro-F1 of 1.000, offering no basis for architecture selection. Under Leave-One-Speed-Out (LOSO) evaluation (train on two speeds, test on the held-out speed), performance drops substantially and diverges across models: BiLSTM 0.180, TCN 0.270, 1D-CNN 0.271, and CNN-BiLSTM 0.401. We trace the LOSO gap to the unbalance centrifugal force law F = meω², which makes speed-confounded features unreliable under cross-speed testing. CNN-BiLSTM improves the mean LOSO macro-F1 by 48% relative to the stronger single-module baseline, 1D-CNN. Although CNN-BiLSTM achieves the highest LOSO performance among the evaluated architectures, it still does not surpass the physics-informed LightGBM baseline of 0.487. Therefore, the primary contribution of this work is not to solve cross-speed diagnosis, but to demonstrate that conventional same-speed evaluation substantially overestimates model capability and that LOSO provides a more deployment-relevant benchmark for future algorithm development. Full article

User Behaviour Analytics (UBA) relies heavily on sequential data to detect anomalies such as insider threats. Traditional approaches often model user behaviour as discrete sequences of events using Recurrent Neural Networks (RNNs) like Long Short-Term Memory (LSTM) networks. These methods implicitly treat time steps as uniform, ignoring the irregular time intervals inherent in user logs. In this paper, we present the first application of Neural Controlled Differential Equations (NCDEs) to user behaviour analytics, a class of continuous-time models that naturally handle irregularly-timed event data. We compare a simple LSTM predictor against an NCDE predictor on the CERT 4.2 and 6.2 insider threat dataset. We demonstrate that standard discrete-time models (LSTMs) produce noisy loss signals on sparse data, forcing downstream classifiers to rely on fragile error spikes. In contrast, Neural CDEs generate stable, continuous error signals. NCDE roughly tripled the F1 of the discrete baseline (0.364 vs. 0.133) on the challenging CERT 6.2 dataset. Full article

Predictive maintenance in underground mining faces challenges due to severe conditions such as confined environments, high humidity, presence of silica dust, and restricted access. This study develops a predictive framework based on oil analysis and machine learning for multiple compartments of mining equipment (engine, hydraulic system, transmission, differential). Samples were processed under ASTM standards, integrating wear metal concentrations (Fe, Cu, Cr, Pb, Al), physicochemical properties (viscosity, TBN, soot), and contaminants (Si, Na). Based on tribology, interpretable ratios were constructed. Three algorithms (Random Forest, Gradient Boosting, and XGBoost) were evaluated using cross-validation. XGBoost achieved the best balance (F1 = 0.852, AUC = 0.975), with a recall of 94.5% for the critical class and only 3 false negatives out of 199 test samples, while Random Forest presented the highest global discrimination power (AUC = 0.978). SHAP revealed that viscosity at 100 °C is the most important predictor (SHAP ~0.9), surpassing iron. No temporal wear trend was found (R² = 0.000). Threshold optimization to 0.25 reduced false negatives by 67% (from 9 to 3). The framework provides interpretable predictions with uncertainty quantification for underground environments. Full article

Unauthorised FM broadcasting poses significant challenges to spectrum regulators globally, contributing to interference, degraded service quality, and national security threats. While traditional spectrum monitoring relies primarily on carrier frequency and power measurements, this study demonstrates that FM baseband features—specifically the multiplex (MPX) signal structure, pilot tone, and Radio Data System (RDS) subcarrier—provide robust discriminative markers for detecting non-compliant transmissions. Using a real-world dataset of 3710 pre-processed records collected across Nigeria’s capital region between 2021 and 2024, we extracted and analysed six transmission parameters: assigned frequency, band occupancy (±100 kHz), MPX overshoot percentage, pilot tone presence, and RDS indicators. A Support Vector Machine (SVM) classifier with radial basis function (RBF) kernel was trained to distinguish compliant licensed stations from regulatory non-compliant transmissions—encompassing both unlicensed transmitters and technically non-compliant licensed operators—achieving 99.96% accuracy, 99.38% precision, and 99.63% recall with a false alarm rate of 0.026%. A Comparative analysis against baseline feature sets confirmed that integrating MPX, pilot, and RDS significantly improved detection robustness compared with carrier-only approaches. Results demonstrate that feature-rich baseband analysis enables scalable, cost-effective regulatory enforcement, reducing manual monitoring burden while enhancing detection reliability. This framework offers practical applicability for spectrum management agencies in resource-constrained environments where unauthorised broadcasting remains prevalent. Full article

Speech-based stress analysis is relevant to forensic-oriented speech processing, security screening, and behavioral monitoring, yet its reliability is often limited by speaker variability, recording conditions, and acoustic mismatch. This study proposes an interpretable multi-domain signal processing framework that models stress-related speech variation through excitation dynamics, vocal tract characteristics, and temporal modulation patterns. The framework integrates source–filter decomposition, residual-domain analysis, harmonic structure analysis, modulation spectrum characterization, and prosodic variability into a unified representation. The SUSAS corpus is used as the primary dataset for supervised stress evaluation. RAVDESS and SAVEE are employed only as controlled arousal-related proxy datasets to examine the consistency of stress-related acoustic patterns, rather than as physiological stress ground truth. VoxCeleb is used exclusively for robustness and domain-variability analysis because it lacks stress labels. For probabilistic evidence assessment, Gaussian mixture models are adopted as the more interpretable density estimator, while normalizing flow is included as a flexible performance-oriented comparator for modeling non-Gaussian feature distributions. Evaluation incorporates likelihood ratio analysis, DET curves, EER, ablation studies, and robustness testing. The proposed framework achieves an EER of 5.8% in the primary supervised evaluation, showing competitive performance while preserving physically meaningful interpretation. Full article

Distributed seismic sensor networks integrated into the Internet of Things (IoT) infrastructure enable continuous condition monitoring of large-scale engineering structures. During long-term operation, however, measurement channels are subject to sensitivity drift, increased noise, and pulse artifacts that statistically mimic real vibration events. Related deep-learning techniques for noisy and ill-posed inverse problems have demonstrated the value of combining principled physical priors with deep models. Although the application domain differs, the underlying methodological insight—that constrained, physics-aware feature mappings can stabilize learning under noisy and partially observed conditions—directly motivates the use of a parameterized quantum circuit as a nonlinear feature transformer in the present work, where Hilbert space mapping serves as an analogous structural prior for the latent representation. Three principal fault modes are considered in this work, corresponding to the dominant degradation mechanisms observed in long-term seismic instrumentation: sensor drift, increased noise, and sensor failure. Each fault mode produces a distinct signature in the windowed feature space; the proposed model is trained to discriminate between them based on the latent CNN-LSTM-VQC representation. We propose a hybrid quantum-inspired deep-learning model (QC-DL) for the detection and diagnosis of channel-degradation anomalies. The architecture combines a 1D-CNN+LSTM feature extractor with a parameterized variational quantum circuit (VQC) used as a nonlinear feature transformer. All quantum experiments were performed on the QPanda3 CPUQVM simulator. The data were split chronologically prior to windowing to avoid information leakage. On real-world labeled accelerometric data with four operating modes (normal/drift/high-noise/failure), the QC-DL model achieved a macro-averaged F1 score of approximately 0.69 and per-class AUC values in the range 0.88–0.99. The mean early-detection latency was 1.6 s versus 2.1 s for the CNN-LSTM baseline (~24% reduction). An ablation study against a parameter-matched classical MLP showed that the gain is modest and not solely attributable to additional nonlinearity. The reported p-values (p = 0.70, p = 0.29) do not establish statistical significance. The results support the feasibility of hybrid quantum-inspired deep learning for sensor-channel verification, while highlighting the need for evaluation on real NISQ hardware. This paper proposes a hybrid quantum-inspired approach for detecting and diagnosing such anomalies in the time series of distributed seismic networks. The architecture combines a classical temporal feature extraction module based on one-dimensional convolutional layers and a recurrent long short-term memory (LSTM) network, which generates a latent window representation of the signal, with a parameterized variational quantum circuit used as a nonlinear feature processor in a hybrid computational circuit. Experimental validation was performed on real-world labeled data with multiple sensor degradation modes. The evaluation was organized in a scoring framework aligned with autonomous operation through window ranking and threshold alarm generation. In the experiments, the proposed model provided a macro-averaged F1 score of approximately 0.69 and area under the receiver operating characteristic (AUC) curve values in the range of 0.88–0.99 across classes, outperforming baseline deep models. The average early detection latency was 1.6 s versus 2.1 s for the baseline recurrent model (a 24% reduction). An ablative comparison with a control model based on a classical multilayer perceptron of comparable dimension confirmed that the improvement is not limited to the addition of additional nonlinearity. The obtained results indicate the potential of quantum-supported deep learning for improving the reliability of long-term vibration monitoring and verifying the correctness of sensor channels in distributed seismic networks. Full article

Manual arc welding is highly dependent on operator skill, leading to variability in weld quality and an increased risk of defects; therefore, reliable monitoring methods for Shielded Metal Arc Welding (SMAW) are required, particularly in manual environments where process variability and environmental noise are inherent. This study proposes a monitoring approach for classifying SMAW process parameters using airborne acoustic signals generated by the welding arc. Welding experiments were conducted on carbon steel plates of different thicknesses (3, 6, and 12 mm) using E6010, E6011, E6013, and E7018 electrodes under Alternating Current (AC) and Direct Current (DC) configurations; acoustic signals were recorded in real time and processed using Mel-Frequency Cepstral Coefficients (MFCCs) and deep audio embeddings from pre-trained VGGish and YAMNet models as inputs to artificial neural network classifiers for multi-output classification of welding process parameters. Model performance was evaluated using per-target metrics (accuracy and macro F1-score) and joint multi-output metrics (Exact Match and Hamming Accuracy). MFCC-based models significantly outperformed embedding-based approaches, achieving up to 94.51% Exact Match and 97.88% Hamming Accuracy, while reducing computational costs. These results demonstrate the feasibility of SMAW monitoring using arc sound, suggesting that spectral features are an effective solution for welding-process monitoring and a promising foundation for future weld-quality monitoring systems. Full article

Wafer and FPGA testing remains costly in semiconductor manufacturing. This paper studies random sampling, stratified sampling, and k-means sampling under a partial-measurement setting with Gaussian Process Regression (GPR), and introduces Short Distance Elimination (SDE), a spatial screening rule that spreads selected training points over the layout. Combining value-based sampling with SDE yields two hybrid methods: S-SDE, which applies SDE within stratified subsets, and K-SDE, which applies SDE within k-means clusters. A calibration-based protocol fixes the value-group labels and SDE thresholds before target-file prediction. The SDE thresholds are selected from $(\alpha ,\beta )$ configurations in $\{0,1,2,3,4\}$, excluding $(0,0)$, using local window statistics and RMSD-based calibration sweeps. Both wafer and FPGA calibration files support $(\alpha ,\beta )=(2,2)$, which is then fixed for all remaining experiments. Experiments on industrial wafer data and actual FPGA silicon data, repeated across four random seeds, show that K-SDE reduces RMSD by 15.84% for wafer data and 13.03% for FPGA data relative to k-means sampling, while S-SDE reduces RMSD by 16.26% and 8.63% relative to stratified sampling. Full article

Automatic modulation classification (AMC) is a core capability for spectrum monitoring, adaptive receivers, and electronic support. Most radio-frequency machine learning (RFML) studies train multi-class classifiers on benchmark datasets that contain a single modulation per recording at baseband. In operational settings, however, the objective is often to detect only a small set of signals of interest, making large multi-class models unnecessarily expensive to train and deploy. In addition, multi-class formulations can increase false-alarm risk due to confusion among non-essential classes and may allocate model capacity inefficiently to distinctions that are irrelevant for the operational objective. This paper investigates an alternative workflow based on targeted binary transformer detectors and evaluates their robustness under practical RF complications. Using the RadioML 2018.01A dataset, we construct binary detection tasks with BPSK as the signal of interest and introduce three increasingly realistic conditions: (i) center-frequency shifts away from baseband, (ii) sampling-rate mismatches via decimation and interpolation, and (iii) multi-signal mixtures where modulations co-occur either in frequency (simultaneous transmissions) or in time (temporal concatenation). The results show that baseband-trained detectors do not generalize to center-frequency-shifted signals, and multi-signal interference can cause complete detection failure unless explicitly modeled during training. We investigate early-exit transformer inference to reduce computation on high-confidence examples, showing it maintains (and occasionally improves) detection performance. We also evaluate inter-modulation transfer learning and intra-modulation adaptation from baseband to mixed- and multi-signal scenarios. Full article

Similar to a web search engine, we have developed a computer-based acoustic search engine tailored for the critical scenario of high-speed aircraft ground explosion monitoring, addressing the long-standing challenge of real-time localization for such high-impact events. Unlike conventional acoustic source localization techniques, our method uniquely resolves the separation and localization of multiple aliasing events, which are prevalent in high-speed aircraft explosion scenarios due to complex shock wave propagation and overlapping signatures. We first calculate the waveforms of all possible acoustic sources over 2D grids. Then, a dimensionality reduction method and fast search technology are applied to the database. Once a high-speed aircraft ground explosion occurs, the real-time system returns detection feedback by matching real-time data with the pre-established search database. Different from other artificial intelligence (AI)-based approaches, the acoustic search engine can handle multiple aliased acoustic events in real time and does not require any prior information or input parameters—a key advantage for emergency response to high-speed aircraft explosions where predefined parameters are often unavailable. Both synthetic tests and field data applications (using actual acoustic records from high-speed aircraft ground explosion experiments) demonstrate the method’s credibility in detecting and localizing multiple acoustic sources. Full article

Speech emotion recognition systems are typically trained on audio sampled at conventional bandwidths that exclude frequencies above approximately 8 kHz, yet the contribution of extended high-frequency information to vocal emotion recognition remains unclear. This study examines how spectral bandwidth influences automatic emotion classification using the RAVDESS corpus of acted speech and song. Recordings were low-pass filtered to simulate multiple bandwidth conditions (8, 12, and 16 kHz, along with the original full-bandwidth signal), and classification was performed using a Random Forest model trained on mel-spectral features. In addition to classification accuracy, we analyzed permutation-based spectral feature importance and the geometry of the classifier’s posterior-probability space. Bandwidth restriction had relatively modest effects on classification accuracy overall, with mean accuracy ranging from approximately 55% to 77% across conditions, although its impact was greater for speech than for song. Feature-importance analyses indicated that the model depends primarily on low- and mid-frequency spectral information, whereas higher-frequency and EHF regions show increased importance when available. Geometry analyses showed no reliable evidence that bandwidth altered the global structure of the stimulus-level emotion space, although spectral truncation reduced separability for certain emotion contrasts, particularly in speech at normal emotional intensity. These results indicate that most acoustic information supporting categorical emotion recognition resides in lower spectral regions, while EHF information provides supplementary acoustic information that may refine some emotional distinctions under specific conditions. Full article