Search Results (1,758)

Accurate detection of plant leaf diseases is essential for enhancing crop productivity and supporting global food security. In addition to disease classification, understanding how environmental and soil conditions affect model performance is important for developing robust real-world agricultural monitoring systems. Although deep learning (DL) models achieve high accuracy on benchmark datasets, their performance in real-world settings is often limited by variations in illumination, background complexity, and environmental conditions. This study proposes a smart DL framework for detecting and classifying multiple leaf diseases in tomato, potato, and pepper plants. The framework combines U²-Net-based leaf segmentation with a Convolutional Neural Network–Bidirectional Gated Recurrent Unit (CNN–Bi-GRU) architecture. MobileNetV2 is employed as the feature extraction backbone to capture spatial characteristics, while Bi-GRU layers model sequential feature dependencies, forming a spatio-temporal network whose architectural design prioritizes parameter efficiency through depthwise separable convolutions and reduced gating complexity. The model was trained and validated using the PlantVillage benchmark dataset and achieved a classification accuracy of 99.8% with a macro-averaged F1-score of 94%, outperforming several state-of-the-art architectures. To assess robustness under real-world conditions, the trained model was further tested on leaf images collected from open-field environments near Eluru, South India. The field evaluation revealed a reduction in classification accuracy to 61.97%, indicating the impact of domain shift and environmental variability. To investigate potential contributing factors, soil parameters, including pH, temperature, moisture, and NPK levels, were monitored using an IoT-based Arduino sensing system over ten consecutive days. Rather than serving as direct inputs to the disease classification model, these environmental measurements were analyzed to assess their potential influence on disease symptom expression and the observed reduction in model performance under field conditions. The results suggest that environmental conditions may influence disease symptom expression and model transferability. This study highlights the importance of integrating DL-based disease recognition with environmental monitoring for reliable field-level agricultural applications. Nevertheless, computational complexity metrics, including inference latency and memory footprint, were not evaluated in the present work and are identified as a priority for future edge deployment studies. Full article

Accurate prediction of natural gas thermophysical properties is essential for applications in production and transportation engineering, including reservoir simulation and flow modeling. Although machine learning (ML) techniques have been widely used, most studies focus on the estimation of these properties, with limited integration into practical applications. In this study, we propose a supervised model based on a Backpropagation Neural Network for simultaneous estimation of four interdependent properties: compressibility factor (Z), viscosity (μ), density (ρ) and gas formation volume factor (B_g). The multi-output model was trained on 58,165 data points generated from thermodynamic correlations, using pressure, temperature, composition (mole fractions of N₂, CO₂ and H₂S), and gas specific gravity as inputs. The results yielded RMSE values of 5.56 × 10⁻⁴, 3.24 × 10⁻⁵, 3.01 × 10⁻², and 6.33 × 10⁻⁴ for Z, μ, ρ and B_g, respectively, with R² coefficients close to unity. The model’s applicability was evaluated by integrating the Z-factor into pressure drop calculations in pipelines using the Cullender and Smith method, resulting in a mean percentage error of 3.78%, close to the traditional method (3.83%). The results indicate that the model is an efficient and consistent alternative, highlighting the potential for integrating ML with classical hydraulic models. Full article

Running neural networks on battery-powered Internet of Things (IoT) sensor nodes is difficult because flash memory, SRAM, latency, and energy per inference are limited at the same time. Existing TinyML co-design methods usually improve model size or memory use, but runtime voltage–frequency control is often handled as a separate step. This separation limits energy saving because the power policy does not use the layer-wise compute profile of the final compressed model. We propose the Cross-Layer Resource Optimizer (CLRO), a three-stage resource optimization pipeline for TinyML inference on an ARM Cortex-M7 target. The first stage, Mixed-Precision Aware Pruning and Distillation (MPAD), assigns per-layer bit widths and pruning ratios using calibration-set sensitivity scores. The second stage, consisting of the Activation Lifetime-Aware Tensor Scheduler (ALTS), uses the compressed graph to find an execution order that reduces peak live static random-access memory (SRAM). The third stage, Reinforcement Learning-Based Dynamic Voltage and Frequency Scaling (DVFS-RL), trains a tabular Q-learning policy from the multiply–accumulate (MAC) utilization profile of the compressed and scheduled model. The learned voltage–frequency policy is stored as a small flash lookup table, so it adds no runtime decision cost during inference. We evaluate the CLRO on all four MLPerf Tiny tasks using an STM32H743ZI microcontroller with 512 kB SRAM and 2 MB flash. The CLRO reaches 91.7% image classification accuracy, 95.4% keyword-spotting accuracy, 89.6% visual wake words accuracy, and 0.913 anomaly detection AUC. The final deployment uses 198 kB flash and 174 kB peak SRAM, with 387 μJ energy per inference and 38 ms latency. Compared with the MCUNet baseline, the CLRO reduces energy by 58.1% and peak SRAM by 39% while keeping the same accuracy level. Full article

Background/Objectives: Multi-leaf collimator (MLC) positioning accuracy critically affects delivered dose fidelity in volumetric modulated arc therapy (VMAT), yet conventional gamma-based quality assurance (QA) provides only plan-level pass/fail outcomes without leaf-specific error localization. This study developed and validated a convolutional neural network (CNN) framework that classifies the magnitude and direction of individual MLC leaf positioning errors directly from fluence map data. Methods: Three patient cohorts were analyzed: 20 prostate cancer patients for model development under an 8:1:1 train/validation/test split and 20 additional prostate and 10 head and neck (H&N) patients reserved for external validation. For inner MLC leaves 21–40, systematic offsets from −5 mm to +5 mm in 1.0 mm increments were independently applied to the two leaf banks, yielding 121 error combinations per leaf. A CNN was trained as a 121-class classifier on two-channel inputs pairing the reference and error-induced fluence map regions and was compared against three tree-based baselines using five-fold cross-validation. Results: The CNN achieved 97.00% accuracy on the internal test set and 96.54 ± 0.43% accuracy across the five patient-level cross-validation folds. Across all test samples, 99.88% and 99.83% of predictions were within 1 mm of the true offset for Bank A and Bank B, respectively, well within the AAPM TG-142 1 mm MLC positioning tolerance. External validation yielded 96.19% accuracy on the additional prostate cohort and 93.72% on the H&N cohort, suggesting reproducibility within the same anatomical site and potential robustness across anatomically distinct treatment sites within a single-institution dataset. Conclusions: The proposed CNN framework demonstrates the feasibility of leaf-specific identification of MLC positioning errors in both magnitude and direction from simulated fluence maps. These findings support further investigation using physically measured fluence data for future clinical translation. Full article

Accurate prediction of significant wave height (SWH) is essential for fisheries management, coastal socioeconomic activities, and marine ecological conservation. In recent years, deep learning-based bias correction has shown considerable potential for improving numerical wave forecasts. However, many existing approaches are still constrained by limited receptive fields and often struggle to capture long-range spatiotemporal dependencies in wave forecast errors. To deal with this issue, we adapt and improve a video prediction framework, namely the Vision Mamba Recurrent Neural Network (VMRNN), to model and correct the spatiotemporal patterns of SWH prediction biases. Comprehensive evaluations show that the multi-channel VMRNN achieves consistently high predictive accuracy across different forecast lead times and sea-state conditions. When validated against reanalysis data, the proposed model reduces the root mean square error (RMSE) of WAVEWATCH III forecasts by 28.2%, 26.1%, and 24.7% at lead times of 24, 48, and 72 h, respectively. It also preserves the spatial structure of SWH fields quite well, with the spatial structural similarity index remaining as high as 0.945 even at the 72 h lead time. Regional assessments over high-wave areas further indicate that VMRNN can effectively reduce both the mean error and the systematic overestimation commonly found in numerical wave models. Additional validation using in situ buoys observations confirms that the model has a robust ability to correct systematic positive biases, especially for wave heights ranging from 0.5 m to 2 m. Taken together, these results suggest that VMRNN has strong spatiotemporal modeling capability and can serve as a promising post-processing framework for improving operational physics-based wave forecasting systems. Full article

This study reports cellulase production by Bacillus paralicheniformis using Plumeria alba leaf powder under submerged fermentation with a focus on systematic bioprocess optimization. Physical parameters were first optimized using a one-factor-at-a-time (OFAT) approach, followed by optimization of yeast extract, MgSO₄ and (NH₄)₂SO₄ via a central composite design (CCD) and response surface methodology (RSM). An artificial neural network (ANN) with a 5:3:1 network trained by the Levenberg–Marquardt algorithm further improved prediction of carboxylmethylcellulase (CMCase) and filter paper cellulase (FPase) activities. This study is the first to exploit Plumeria alba leaf powder as an untapped, low-cost lignocellulosic substrate for cellulase production by B. paralicheniformis and uniquely benchmarks RSM against ANN-based modeling to identify superior predictive frameworks for bioprocess optimization. Under optimized conditions (24 h, 4% w/v substrate, 1% v/v inoculum), the maximum FPase and CMCase activities reached 60.53 IU/mL/min and 332.10 IU/mL/min respectively. Partial characterization showed optimum FPase and CMCase activities at 50 °C and 70 °C, respectively, at pH 7.5. Enzymes also showed activation by NaCl and some select solvents while tolerating a broad range of metal ions. The enzymatic hydrolysis of P. alba biomass released 59.42 mg/mL total reducing sugars after 8hr, confirming efficient saccharification from a low-cost feedstock. The ANN model (R² = 97.59% for CMCase; 85.95% for FPase) outperformed RSM (R² = 85.95% and 78.25%, respectively), while radial basis function optimization reached 99.99%. These findings highlight B. paralicheniforms cellulase as a promising biocatalyst for biorefinery applications and demonstrate the value of integrating RSM and ANN for process optimization. Full article

Foundation time-series models have recently shown potential for forecasting complex temporal signals, but their behavior in patient-specific continuous glucose monitoring (CGM) forecasting remains insufficiently understood, particularly when only glucose history is available. This study provides a patient-level benchmark of foundation models for 30 min ahead glucose prediction in adults with type 1 diabetes mellitus (T1DM) under a strictly univariate CGM-only setting. Using the HUPA–UCM dataset from 25 individuals, we evaluated TimeGPT, Chronos, and Sundial against representative statistical, machine learning, and deep learning forecasters, including ARIMA, ETS, gradient-boosting models, recurrent networks, and neural forecasting architectures. Models were assessed using a local walk-forward validation strategy over the final 24 h of CGM data for each patient. Foundation models achieved the strongest global performance, with Sundial obtaining the lowest overall MAE ($6.06\mathrm{mg}/\mathrm{dL}$), while TimeGPT and Chronos remained among the most competitive approaches. However, patient-level analyses showed that this advantage was not uniform: ARIMA remained highly competitive in selected individuals, and no single model consistently dominated across the cohort. These findings suggest that foundation time-series models are promising tools for short-horizon CGM forecasting, but their use should be framed within patient-specific model selection rather than as universal replacements for classical forecasting methods. Full article

The ongoing global energy transition has intensified the need for precise modeling of renewable energy systems, especially photovoltaic–thermal (PV/T) systems that have the ability to produce both electrical and thermal energy. Improving the efficiency and reliability of PV/T systems is a key enabler of the transition toward sustainable energy. Accurate forecasting of their thermal performance is therefore essential to maximize renewable energy use and reduce energy losses. A deep learning-based method is proposed in this study for the prediction of the thermal efficiency of an air-based PV/T system. More specifically, temporal deep learning architectures are investigated to exploit the complex nonlinear relationships and temporal dependencies governing the thermal behavior of the PV/T collector. A comprehensive comparative analysis is conducted using four state-of-the-art architectures, namely Temporal Convolutional Network (TCN), Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), and Transformer. Furthermore, the influence of sequence length is examined through a sensitivity analysis considering forecasting horizons of 1 h, 6 h, 12 h, and 24 h. The models are evaluated using the Coefficient of Determination (R²), Root Mean Square Error (RMSE), and Mean Absolute Error (MAE). The results demonstrate that forecasting performance is strongly influenced by the selected temporal horizon. Among the investigated configurations, the 24-h horizon provided the most informative temporal context for thermal efficiency prediction. Under this common forecasting horizon, the LSTM model achieved the highest predictive accuracy, reaching an R² of 0.9952, an RMSE of 0.5975, and an MAE of 0.2364, outperforming the TCN, GRU, and Transformer architectures. The residual error and convergence analyses further highlighted the effectiveness of recurrent neural networks in capturing the thermal dynamics of the investigated PV/T system. By enabling accurate and reliable thermal efficiency forecasting, the proposed framework supports improved energy management, higher energy efficiency, and a stronger integration of renewable energy systems, thus contributing to more sustainable operation of hybrid solar energy technologies. Full article

Decoding short-window electroencephalography (EEG) signals is critical for low-latency brain–computer interfaces (BCIs), yet current models struggle to extract robust features under high cross-subject variability and low signal-to-noise ratios. To address this, we propose a spatiotemporal decoding framework integrating dynamic weighted permutation entropy (DWPE) with a hybrid neural network. We introduce DWPE to quantify nonlinear dynamic complexity while retaining amplitude information. These features are subsequently processed by a cascaded convolutional neural network (CNN) and bidirectional long short-term memory (BiLSTM) architecture with spatial attention, enabling the simultaneous extraction of topological patterns and temporal dependencies. The framework was evaluated on three public motor imagery datasets (hBCI, BCI Competition IV-2a, and IV-2b) using a fixed 3 s window. Empirical results demonstrate that our approach achieves an average accuracy of 84.35% and an AUC of 0.8821 on the hBCI dataset, significantly outperforming current representative recent baselines (p < 0.01). Ablation studies confirm that integrating DWPE yields a 3.89% accuracy improvement over the spatial–temporal backbone alone. With a single-sample inference time of 20.94 ms and an estimated total decision latency of approximately 3.02 s under the 3 s window setting, the proposed method provides a favorable balance between decoding accuracy and computational efficiency for short-window and near-online BCI applications. Full article

Background/Objectives: Reliable detection of epileptic seizures using electroencephalography (EEG) is crucial for clinical diagnosis and for alleviating clinicians’ workload. However, existing studies still make insufficient use of phase information, and the synergy between local time–frequency pattern extraction and global dependency modeling remains limited. Methods: We propose a seizure detection framework based on the continuous wavelet transform (CWT), a three-dimensional convolutional neural network (3D-CNN), and a vision transformer (ViT). First, multichannel EEG segments are preprocessed, after which CWT is used to generate power spectrograms and phase spectrograms. These representations are then fused along the depth dimension into a unified power-phase volume and fed into a hybrid network composed of a 3D-CNN feature extractor and a single-layer ViT encoder to jointly learn local time–frequency–channel coupling patterns and higher-level global dependencies. Finally, seizure detection is completed by combining moving-average filtering, thresholding, and collar correction. Results: On the public CHB-MIT dataset and the clinical SH-SDU dataset, the proposed method achieved average segment-level sensitivities of 98.68% and 92.05%, specificities of 98.33% and 97.53%, accuracies of 98.49% and 96.37%, and AUC values of 97.26% and 92.89%, respectively. In event-level evaluation, the average sensitivities were 99.13% and 96.08%, with false detection rates of 0.88/h and 0.69/h, respectively. Further multi-stage ablation experiments together with t-SNE and Grad-CAM visualizations provided qualitative and experimental support for the design rationale of the joint power-phase input and the hybrid 3D-CNN-ViT architecture. Conclusions: The proposed framework effectively exploits the complementary discriminative value of power and phase information in epileptic EEG and demonstrates strong detection performance under patient-specific evaluation on both public and clinically collected datasets. Full article

Road quality monitoring is necessary for safety, ride comfort, and driver-assistance systems. The knowledge of road features enables preventive and corrective actions at vehicle and infrastructure levels. While deep learning models are effective for surface classification, transitioning them to real-time embedded environments requires optimization. This study deploys a model based on convolutional and long short-term memory neural networks to classify five road conditions using continuous vibration data from multiple inertial measurement units. Executed on a MicroAutoBox III Embedded PC, the system preprocesses data at vehicle speeds between 5.0 and 25.0 km/h. Compared to the offline baseline deployment, this edge-optimized architecture reduced inference latency by 88% (from 33.8 ms to 4.05 ms) while maintaining a fair weighted-average F1-score of 0.8751 in real-world, cross-platform conditions (against the offline baseline average F1-score of 0.9338). This processing time operates within the 11.6 ms limit required by the 86 Hz sensor polling rate. Additionally, geospatial mapping was able to localize structural anomalies, showing robustness to environmental lighting conditions, which frequently affect vision-based systems. This cyber-physical deployment suggests the feasibility of executing temporal deep learning real-time models. Future work will target highway-speed validation and domain adaptation to assess transferability across diverse vehicle suspensions. Full article

Although coconut water is recognized for its desirable sensory appeal and nutritional composition, its broader industrial use is constrained by the rapid deterioration that occurs after extraction. In this study, crossflow microfiltration of coconut water with a silicon carbide membrane was optimized by investigating pressure and temperature through a face-centered design (FCD) and artificial neural network modeling coupled with a genetic algorithm (ANN–GA). Permeate flux and fouling index were used as process responses, and the optimized condition was further examined in terms of hydraulic resistance, fouling behavior, and retention of minerals and primary metabolites. Pressure and temperature affected the process differently: permeate flux showed marked nonlinear behavior, whereas fouling index was governed mainly by pressure. At the sample level, ANN described permeate flux more accurately than FCD (R² = 0.99 vs. 0.96), whereas FCD showed better grouped cross-validated predictivity across unseen pressure–temperature conditions (Q² = 0.85 vs. 0.57). For the fouling index, FCD outperformed ANN in both sample-level fit and grouped validation (R² = 0.95 vs. 0.60; Q² = 0.70 vs. 0.61). Both approaches converged on the same favorable operating window, and experimental validation at 60 kPa and 35 °C yielded 1085.23 ± 23.12 L h⁻¹ m⁻² and 83.56 ± 1.56%. During concentration mode, flux decline was severe but predominantly reversible, with high clean-water permeance recovery after chemical cleaning. Resistance partition and fouling modeling indicated that the main hydraulic limitation was associated with concentration polarization and external cake-layer buildup rather than irreversible membrane damage. The clarified fraction also preserved high transmission of major minerals and relevant primary metabolites, indicating that the selected condition combined high productivity, manageable fouling, and satisfactory nutritional retention. Full article

This research proposes a physics-based deep learning framework, developed as a proof-of-concept based on synthetic data, for estimating the mechanical properties of layered media—namely density (ρ), Young’s modulus (E), and top layer thickness (h₁)—using synthetic seismogram images generated via Finite Element Method (FEM) simulations. The dataset, comprising 5000 simulations, incorporates physical constraints and empirical density–modulus correlations. While a ResNet-style Convolutional Neural Network (CNN) extracts density and stiffness parameters from composite time–frequency images, the estimation of h1 utilizes a direct time-domain raw-signal approach to preserve spatial resolution. A 5-fold nested cross-validation scheme with internal Bayesian Optimization ensures rigorous model evaluation, further validated by normality assessments and bootstrap confidence intervals. Performance was tested against synthetic Gaussian noise (0% to 50%) and benchmarked against classical Full Waveform Inversion (FWI). The results demonstrate high predictive accuracy for shallow properties, with R² values reaching 0.96 for Young’s modulus and 0.83 for raw-signal thickness. The neural network model requires 0.035 s per inference compared to 180 s for the FWI approach, avoiding the local minima convergence issues typical of iterative inversion. The framework exhibits resilience under moderate noise levels (up to 30%), establishing a reliable baseline for future experimental validation. Full article

Predicting the wind field around military vehicles during extended missions is crucial to avoid detectability by infrared (IR) devices. This is a challenging task because of the geometric complexity of the vehicles and the unpredictable nature of wind direction, which can shift abruptly and have a significant impact on the flow field and heat transfer. Computational fluid dynamics (CFD) is routinely used to calculate flow fields around ground vehicles. However, this requires extensive computational time and memory, making it unsuitable for real-time analysis. To address these challenges, this paper focuses on machine learning (ML) techniques for accurate wind field prediction in real time for unseen wind directions within the sampled range. Reduced order modeling (ROM) is used for dimensionality reduction of flow field data derived from high-fidelity CFD simulations. ML models are trained using low-dimensional data from the ROM, and the predicted low-dimensional data for unseen wind directions by the trained ML model is used to reconstruct the flow field. ROM, in conjunction with ML techniques, offers a substantial reduction in analysis time while maintaining the ability to predict the flow field accurately. In this study, a neural network architecture with three output formulations trained using ROM data was used for the predictions, and the accuracy of the formulations was evaluated by comparing them with the CFD results. An optimal ML model is identified by varying the number of hidden layers and neurons within those layers. The developed ROM- and ML-based approach was able to predict the unseen flow field in less than a second, while a single CFD simulation required approximately 2.6 h per wind direction. Full article

If concentrating coal is difficult or impossible using gravity methods (such as jigs, shaking spirals, dense-media drum, and cyclone), which are among the cheapest and simplest options, flotation becomes an alternative. This is due to the differences in surface chemistry properties between the relatively hydrophobic coal and the gangue minerals. On the other hand, flotation methods are far more complex than gravity methods and involve many more parameters that influence concentrate, such as coal particle size, amounts of reagents dosages (e.g., collectors, activators, depressants, and frothers), conditioning times, pulp mixing speeds, flotation times, and pH levels of the pulp medium. In flotation methods with so many variables, determining the combustible recovery (CR) and ash content (AC) of clean coal concentrate that can be obtained may require many experiments. To facilitate these challenging processes, understand the effects of parameters influencing concentration on the flotation method, and estimate the resulting clean coal recovery and ash content, it is necessary to utilize various statistical regression methods. In this study, the effects of six parameters on the flotation of a lignite coal sample with 40% ash content were used to estimate the CR and AC of coal concentrate using multivariate linear regression (MLR) and artificial neural network (ANN) models. As a result, the ANN model demonstrated superior estimate accuracy, with correlation coefficients of 0.988 and 0.963, compared with the MLR models (R² = 0.575 and 0.540) for estimating the ash content (AC, %) and combustible recovery (CR, %) of coal concentrate, respectively. Full article