Search Results (268)

The Denoising Wavelet Transform Convolutional Long Short-Term Memory Network (DWTConvLSTM) is a novel ionospheric total electron content (TEC) spatiotemporal prediction model proposed in 2025 that can simultaneously consider high-frequency and low-frequency features while suppressing noise. However, it also has flaws as it only considers unidirectional temporal features in spatiotemporal prediction. To address this issue, this paper adopts a bidirectional structure and designs a bidirectional DWTConvLSTM model that can simultaneously extract bidirectional spatiotemporal features from TEC maps. Furthermore, we integrate a lightweight attention mechanism called Convolutional Additive Self-Attention (CASA) to enhance important features and attenuate unimportant ones. The final model was named CASA-BiDWTConvLSTM. We validated the effectiveness of each improvement through ablation experiments. Then, a comprehensive comparison was performed on the 11-year Global Ionospheric Maps (GIMs) dataset, involving the proposed CASA-BiDWTConvLSTM model and several other state-of-the-art models such as C1PG, ConvGRU, ConvLSTM, and PredRNN. In this experiment, the dataset was partitioned into 7 years for training, 2 years for validation, and the final 2 years for testing. The experimental results indicate that the RMSE of CASA-BiDWTConvLSTM is lower than those of C1PG, ConvGRU, ConvLSTM, and PredRNN. Specifically, the decreases in RMSE during high solar activity years are 24.84%, 16.57%, 13.50%, and 10.29%, respectively, while the decreases during low solar activity years are 26.11%, 16.83%, 11.68%, and 7.04%, respectively. In addition, this article also verified the effectiveness of CASA-BiDWTConvLSTM from spatial and temporal perspectives, as well as on four geomagnetic storms. Full article

Accurate real-time information about soil moisture (SM) at a large scale is essential for improving hydrological modeling, managing water resources, and monitoring extreme weather events. This study presents a framework using convolutional long short-term memory (ConvLSTM) network to produce short- (1, 3, and 7 days ahead) and mid-term (14 and 30 days ahead) forecasts of SM at surface (0–10 cm) and subsurface (10–40 and 40–100 cm) soil layers across the contiguous U.S. The model was trained with five-year period (2018–2022) datasets including Soil Moisture Active Passive (SMAP) level 3 ancillary covariables, North American Land Data Assimilation System phase 2 (NLDAS-2) SM product, shortwave infrared reflectance from Moderate Resolution Imaging Spectroradiometer (MODIS), and terrain features (e.g., elevation, slope, curvature), as well as soil texture and bulk density maps from the Soil Landscape of the United States (SOLUS100) database. To develop and evaluate the model, the dataset was divided into three subsets: training (January 2018–January 2021), validation (2021), and testing (2022). The outputs were validated with observed in situ data from the Soil Climate Analysis Network (SCAN) and the United States Climate Reference Network (USCRN) soil moisture networks. The results indicated that the accuracy of SM forecasts decreased with increasing lead time, particularly in the surface (0–10 cm) and subsurface (10–40 cm) layers, where strong fluctuations driven by rainfall variability and evapotranspiration fluxes introduced greater uncertainty. Across all soil layers and lead times, the model achieved a median unbiased root mean square error (ubRMSE) of 0.04 cm³ cm⁻³ with a Pearson correlation coefficient of 0.61. Further, the performance of the model was evaluated with respect to both land cover and soil texture databases. Forecast accuracy was highest in coarse-textured soils, followed by medium- and fine-textured soils, likely because the greater penetration depth of microwave observations improves SM retrieval in sandy soils. Among land cover types, performance was strongest in grasslands and savannas and weakest in dense forests and shrublands, where dense vegetation attenuates the microwave signal and reduces SM estimation accuracy. These results demonstrate that the ConvLSTM framework provides skillful short- and mid-term forecasts of surface and subsurface soil moisture, offering valuable support for large-scale drought and flood monitoring. Full article

Background/Objectives: Lung and colon cancers remain among the most prevalent and fatal diseases worldwide, and their early detection is a serious challenge. The data used in this study was obtained from the Lung and Colon Cancer Histopathological Images Dataset, which comprises five different classes of image data, namely colon adenocarcinoma, colon normal, lung adenocarcinoma, lung normal, and lung squamous cell carcinoma, split into training (80%), validation (10%), and test (10%) subsets. In this study, we propose the ViT-DCNN (Vision Transformer with Deformable CNN) model, with the aim of improving cancer detection and classification using medical images. Methods: The combination of the ViT’s self-attention capabilities with deformable convolutions allows for improved feature extraction, while also enabling the model to learn both holistic contextual information as well as fine-grained localized spatial details. Results: On the test set, the model performed remarkably well, with an accuracy of 94.24%, an F1 score of 94.23%, recall of 94.24%, and precision of 94.37%, confirming its robustness in detecting cancerous tissues. Furthermore, our proposed ViT-DCNN model outperforms several state-of-the-art models, including ResNet-152, EfficientNet-B7, SwinTransformer, DenseNet-201, ConvNext, TransUNet, CNN-LSTM, MobileNetV3, and NASNet-A, across all major performance metrics. Conclusions: By using deep learning and advanced image analysis, this model enhances the efficiency of cancer detection, thus representing a valuable tool for radiologists and clinicians. This study demonstrates that the proposed ViT-DCNN model can reduce diagnostic inaccuracies and improve detection efficiency. Future work will focus on dataset enrichment and enhancing the model’s interpretability to evaluate its clinical applicability. This paper demonstrates the promise of artificial-intelligence-driven diagnostic models in transforming lung and colon cancer detection and improving patient diagnosis. Full article

This study proposes a Compressive Sensing Convolutional Long Short-Term Memory (CSCL) model that aims to improve short-term (12–24 h) forecast accuracy compared to standard ConvLSTM. It is especially useful when subtle spatiotemporal variations complicate feature extraction. CSCL uses uniform sampling to partially mask spatiotemporal wave fields. The model training strategy integrates both complete and masked samples from pre- and post-sampling. This design encourages the network to learn and amplify subtle distributional differences. Consequently, small variations in convolutional responses become more informative for feature extraction. We considered the theoretical explanations for why this sampling-augmented training enhances sensitivity to minor signals and validated the approach experimentally. For the region 120–140° E and 20–40° N, a four-layer CSCL model using the first five moments as inputs achieved the best prediction performance. Compared to ConvLSTM, the R² for significant wave height improved by 2.2–43.8% and for mean wave period by 3.7–22.3%. A wave-energy case study confirmed the model’s practicality. CSCL may be extended to the prediction of extreme events (e.g., typhoons, tsunamis) and other oceanic variables such as wind, sea-surface pressure, and temperature. Full article

We compare several methods for predicting future built-up land cover using only a short yearly time series of satellite-derived binary urban maps. Existing methods of built-up expansion forecasting often rely on ancillary datasets such as utility networks, distance to transportation nodes, and population density maps, along with remotely sensed aerial or satellite imagery. Such ancillary datasets are not always available and lack the temporal density of satellite imagery. Moreover, existing work often focuses on quantifying the expected volume of built-up expansion, rather than predicting where exactly that expansion will occur. To address these gaps, we evaluate six methods for the creation of prediction maps showing expected areas of future built-up expansion, using yearly built/not-built maps derived from Sentinel-2 imagery as inputs: Cellular Automata, logistic regression, Support Vector Machines, Random Forests, Convolutional Neural Networks (CNNs), and CNNs with the addition of long short-term memory (ConvLSTM). Of these six, we find CNNs to be the best-performing method, with an average Cohen’s kappa score of 0.73 across nine study sites in the continental United States. Full article

SA-ConvLSTM is a recently proposed spatiotemporal model for total electron content (TEC) prediction, which effectively catches long-term temporal evolution and global-scale spatial correlations in TEC. However, its reliance on standard convolution limits spatial feature extraction to fixed regular regions, reducing the flexibility for irregular TEC variations. To address this limitation, we enhance SA-ConvLSTM by incorporating deformable convolution, proposing SA-DConvLSTM. This achieves adaptive spatial feature extraction through learnable offsets in convolutional kernels. Building on this improvement, we design ED-SA-DConvLSTM, a TEC spatiotemporal prediction model based on an encoder–decoder architecture with SA-DConvLSTM as its fundamental block. Firstly, the effectiveness of the model improvement was verified through an ablation experiment. Subsequently, a comprehensive quantitative comparison was conducted between ED-SA-DConvLSTM and baseline models (C1PG, ConvLSTM, and ConvGRU) in the region of 12.5° S–87.5° N and 25° E–180° E. The experimental results showed that the ED-SA-DConvLSTM exhibited superior performance compared to C1PG, ConvGRU, and ConvLSTM, with prediction accuracy improvements of 10.27%, 7.65%, and 7.16% during high solar activity and 11.46%, 4.75%, and 4.06% during low solar activity, respectively. To further evaluate model performance under extreme conditions, we tested the ED-SA-DConvLSTM during four geomagnetic storms. The results showed that the proportion of its superiority over the baseline models exceeded 58%. Full article

Accurate fire source location in an aircraft cargo compartment cannot be determined by common design practices. This study proposes an advanced fire location inversion framework based on a Convolutional Long-Short-Term Memory (ConvLSTM) network. A self-designed interpolation preprocessing module is introduced to realize the integration of spatial and temporal sensor data. The model was trained and validated using a comprehensive database generated from large-scale fire dynamics simulations. Hyperparameter optimization, including a learning rate of 0.001 and a 5 × 5 convolution kernel size, can effectively avoid the systematic errors introduced by interpolation preprocessing, further enhancing model robustness. Validation in simplified scenarios demonstrated a mean squared error of 0.0042 m and a mean positional deviation of 0.095 m for the fire source location. Moreover, the present study assessed the model’s timeliness and reliability in full-scale cabin complex scenarios. The model maintained high performance across varying heights within cargo compartments, achieving a correlation coefficient of 0.99 and a mean absolute relative error of 1.9%. Noteworthily, reasonable location accuracy can be achieved with a minimum of three detectors, even in obstructed environments. These findings offer a robust tool for enhancing fire safety systems in aviation and other similar complex scenarios. Full article

Despite the pivotal role of unmanned aerial vehicles (UAVs) in intelligent inspection tasks, existing video instance segmentation methods struggle with irregular deforming targets, leading to inconsistent segmentation results due to ineffective feature offset capture and temporal correlation modeling. To address this issue, we propose a hierarchical offset compensation and temporal memory update method for video instance segmentation (HT-VIS) with a high generalization ability. Firstly, a hierarchical offset compensation (HOC) module in the form of a sequential and parallel connection is designed to perform deformable offset for the same flexible target across frames, which benefits from compensating for spatial motion features at the time sequence. Next, the temporal memory update (TMU) module is developed by employing convolutional long-short-term memory (ConvLSTM) between the current and adjacent frames to establish the temporal dynamic context correlation and update the current frame feature effectively. Finally, extensive experimental results demonstrate the superiority of the proposed HDNet method when applied to the public YouTubeVIS-2019 dataset and a self-built UAV-Seg segmentation dataset. On four typical datasets (i.e., Zoo, Street, Vehicle, and Sport) extracted from YoutubeVIS-2019 according to category characteristics, the proposed HT-VIS outperforms the state-of-the-art CNN-based VIS methods CrossVIS by 3.9%, 2.0%, 0.3%, and 3.8% in average segmentation accuracy, respectively. On the self-built UAV-VIS dataset, our HT-VIS with PHOC surpasses the baseline SipMask by 2.1% and achieves the highest average segmentation accuracy of 37.4% in the CNN-based methods, demonstrating the effectiveness and robustness of our proposed framework. Full article

Radar echo extrapolation is a critical forecasting tool in the field of meteorology, playing an especially vital role in nowcasting and weather modification operations. In recent years, spatiotemporal sequence prediction models based on deep learning have garnered significant attention and achieved notable progress in radar echo extrapolation. However, most of these extrapolation network architectures are built upon convolutional neural networks, using radar echo images as input. Typically, radar echo intensity values ranging from −5 to 70 dBZ with a resolution of 5 dBZ are converted into 0–255 grayscale images from pseudo-color representations, which inevitably results in the loss of important echo details. Furthermore, as the extrapolation time increases, the smoothing effect inherent to convolution operations leads to increasingly blurred predictions. To address the algorithmic limitations of deep learning-based echo extrapolation models, this study introduces three major improvements: (1) A Deep Convolutional Generative Adversarial Network (DCGAN) is integrated into the ConvLSTM-based extrapolation model to construct a DCGAN-enhanced architecture, significantly improving the quality of radar echo extrapolation; (2) Considering that the evolution of radar echoes is closely related to the surrounding meteorological environment, the study incorporates specific physical variable products from the initial zero-hour field of RMAPS-NOW (the Rapid-update Multiscale Analysis and Prediction System—NOWcasting subsystem), developed by the Institute of Urban Meteorology, China. These variables are encoded jointly with high-resolution (0.5 dB) radar mosaic data to form multiple radar cells as input. A multi-channel radar echo extrapolation network architecture (MR-DCGAN) is then designed based on the DCGAN framework; (3) Since radar echo decay becomes more prominent over longer extrapolation horizons, this study departs from previous approaches that use a single model to extrapolate 120 min. Instead, it customizes time-specific loss functions for spatiotemporal attenuation correction and independently trains 20 separate models to achieve the full 120 min extrapolation. The dataset consists of radar composite reflectivity mosaics over North China within the range of 116.10–117.50°E and 37.77–38.77°N, collected from June to September during 2018–2022. A total of 39,000 data samples were matched with the initial zero-hour fields from RMAPS-NOW, with 80% (31,200 samples) used for training and 20% (7800 samples) for testing. Based on the ConvLSTM and the proposed MR-DCGAN architecture, 20 extrapolation models were trained using four different input encoding strategies. The models were evaluated using the Critical Success Index (CSI), Probability of Detection (POD), and False Alarm Ratio (FAR). Compared to the baseline ConvLSTM-based extrapolation model without physical variables, the models trained with the MR-DCGAN architecture achieved, on average, 18.59%, 8.76%, and 11.28% higher CSI values, 19.46%, 19.21%, and 19.18% higher POD values, and 19.85%, 11.48%, and 9.88% lower FAR values under the 20 dBZ, 30 dBZ, and 35 dBZ reflectivity thresholds, respectively. Among all tested configurations, the model that incorporated three physical variables—relative humidity (rh), u-wind, and v-wind—demonstrated the best overall performance across various thresholds, with CSI and POD values improving by an average of 16.75% and 24.75%, respectively, and FAR reduced by 15.36%. Moreover, the SSIM of the MR-DCGAN models demonstrates a more gradual decline and maintains higher overall values, indicating superior capability in preserving echo structural features. Meanwhile, the comparative experiments demonstrate that the MR-DCGAN (u, v + rh) model outperforms the MR-ConvLSTM (u, v + rh) model in terms of evaluation metrics. In summary, the model trained with the MR-DCGAN architecture effectively enhances the accuracy of radar echo extrapolation. Full article

High-accuracy streamflow forecasting with long lead times can help promote the efficient utilization of water resources. However, the construction of cascade reservoirs has allowed the evolution of natural continuous rivers into multi-block rivers. The existing streamflow forecasting methods fail to consider the impact of reservoir operation. Thus, a novel short-term streamflow forecasting method for multi-block watersheds was proposed by integrating machine learning and hydrological models. Firstly, based on IMERG precipitation, the forecast precipitation product’s error is corrected by the long short-term memory neural network (LSTM). Secondly, coupling convolutional LSTM (ConvLSTM) and LSTM, operation rules for cascade reservoirs are extracted. Thirdly, a short-term deterministic streamflow forecasting model was built for multi-block watersheds. Finally, according to the sources of forecasting errors, probabilistic streamflow forecasting models based on the Gaussian mixture model (GMM) were proposed, and their performances were compared. Taking the Yalong River as an example, the main results are as follows: (1) Deep learning models (ConvLSTM and LSTM) show good performance in forecast precipitation correction and reservoir operation rule extraction, contributing to streamflow forecasting accuracy. (2) The proposed streamflow deterministic forecasting method has good forecasting performance with NSE above 0.83 for the following 1–5 days. (3) The GMM model, using upstream evolutionary forecasted streamflow, interval forecasted streamflow, and downstream forecasted streamflow as the input–output combination, has good probabilistic forecasting performance and can adequately characterize the “non-normality” and “heteroskedasticity” of forecasting uncertainty. Full article

In this paper, we propose TransConvNet, a hybrid model combining Convolutional Neural Networks (CNNs), self-attention mechanisms, and transfer learning for wireless signal recognition under challenging conditions. The model effectively addresses challenges such as low signal-to-noise ratio (SNR), low sampling rates, and limited labeled data. The CNN module extracts local features and suppresses noise, while the self-attention mechanism within the Transformer encoder captures long-range dependencies in the signal. To enhance performance with limited data, we incorporate transfer learning by leveraging pre-trained models, ensuring faster convergence and improved generalization. Extensive experiments were conducted on a six-class wireless signal dataset, downsampled to 1 MSPS to simulate real-world constraints. The proposed TransConvNet achieved 92.1% accuracy, outperforming baseline models such as LSTM, CNN, and RNN across multiple evaluation metrics, including RMSE and $R^2$. The model demonstrated strong robustness under varying SNR conditions and exhibited superior discriminative ability, as confirmed by Precision–Recall and ROC curves. These results validate the effectiveness and robustness of the TransConvNet model for wireless signal recognition, particularly in resource-constrained and noisy environments. Full article

Accurate traffic flow prediction is essential for highway operations, especially during holidays when surging traffic poses significant challenges. This study focuses on holiday traffic and introduces a spatiotemporal cross-attention network (ST-Cross-Attn) that combines a bidirectional convolutional LSTM (Bi-ConvLSTM) with a cross-attention module to jointly predict toll station inbound flow and outbound flow. Under the multi-task learning framework, the model shares spatial–temporal features between inbound flow and outbound flow, enhancing their representations and improving multi-step prediction accuracy. Using three years of highway traffic flow data during Labor Day from Shandong, China, ST-Cross-Attn outperformed eight state-of-the-art benchmarks, achieving an average improvement of 4.34% in inbound flow prediction and 2.3% in outbound flow prediction. Extensive ablation studies further confirmed the effectiveness of the model’s components and multi-task learning framework, demonstrating its potential for reliable holiday traffic forecasting. Full article

This review critically analyzes recent advancements in dairy cow behavior recognition, highlighting novel methodological contributions through the integration of advanced artificial intelligence (AI) techniques such as transformer models and multi-view tracking with social network analysis (SNA). Such integration offers transformative opportunities for improving dairy cattle welfare, but current applications remain limited. We describe the transition from manual, observer-based assessments to automated, scalable methods using convolutional neural networks (CNNs), spatio-temporal models, and attention mechanisms. Although object detection models, including You Only Look Once (YOLO), EfficientDet, and sequence models, such as Bidirectional Long Short-Term Memory (BiLSTM) and Convolutional Long Short-Term Memory (convLSTM), have improved detection and classification, significant challenges remain, including occlusions, annotation bottlenecks, dataset diversity, and limited generalizability. Existing interaction inference methods rely heavily on distance-based approximations (i.e., assuming that proximity implies social interaction), lacking the semantic depth essential for comprehensive SNA. To address this, we propose innovative methodological intersections such as pose-aware SNA frameworks and multi-camera fusion techniques. Moreover, we explicitly discuss ethical challenges and data governance issues, emphasizing data transparency and animal welfare concerns within precision livestock contexts. We clarify how these methodological innovations directly impact practical farming by enhancing monitoring precision, herd management, and welfare outcomes. Ultimately, this synthesis advocates for strategic, empathetic, and ethically responsible precision dairy farming practices, significantly advancing both dairy cow welfare and operational effectiveness. Full article

The ionospheric total electron content (TEC) has complex spatiotemporal variations, making its spatiotemporal prediction challenging. Capturing long-range spatial dependencies is of great significance for improving the spatiotemporal prediction accuracy of TEC. Existing work based on Convolutional Long Short-Term Memory (ConvLSTM) primarily relies on convolutional operations for spatial feature extraction, which are effective at capturing local spatial correlations, but struggle to model long-range dependencies, limiting their predictive performance. Self-Attention Convolutional Long Short-Term Memory (SA-ConvLSTM) can selectively store and focus on long-range spatial dependencies, but it requires the input length and output length to be the same due to its “n vs. n” structure, limiting its application. To solve this problem, this paper proposes an encoder-decoder SA-ConvLSTM, abbreviated as ED-SA-ConvLSTM. It can effectively capture long-range spatial dependencies using SA-ConvLSTM and achieve unequal input-output lengths through encoder–decoder structure. To verify its performance, the proposed ED-SA-ConvLSTM was compared with C1PG, ConvLSTM, and PredRNN from multiple perspectives in the area of 12.5° S–87.5° N, 25° E–180° E, including overall quantitative comparison, comparison across different months, comparison at different latitude regions, visual comparisons, and comparison under extreme situations. The results have shown that, in the vast majority of cases, the proposed ED-SA-ConvLSTM outperforms the comparative models. Full article

Wind power forecasting is challenging because of complex, nonlinear relationships between inherent patterns and external disturbances. Though much progress has been achieved in deep learning approaches, existing methods cannot effectively decompose and model intertwined spatio-temporal dependencies. Current methods typically treat wind power as a unified signal without explicitly separating inherent patterns from external influences, so they have limited prediction accuracy. This paper introduces a novel framework GCN-EIF that decouples external interference factors (EIFs) from inherent wind power patterns to achieve excellent prediction accuracy. Our innovation lies in the physically informed architecture that explicitly models the mathematical relationship: $P\left(t\right)=P_{\mathrm{inherent}}\left(t\right)+\mathrm{EIF}\left(t\right)$. The framework adopts a three-component architecture consisting of (1) a multi-graph convolutional network using both geographical proximity and power correlation graphs to capture heterogeneous spatial dependencies between wind farms, (2) an attention-enhanced LSTM network that weights temporal features differentially based on their predictive significance, and (3) a specialized Conv2D mechanism to identify and isolate external disturbance patterns. A key methodological contribution is our signal decomposition strategy during the prediction phase, where an EIF is eliminated from historical data to better learn fundamental patterns, and then a predicted EIF is reintroduced for the target period, significantly reducing error propagation. Extensive experiments across diverse wind farm clusters and different weather conditions indicate that GCN-EIF achieves an 18.99% lower RMSE and 5.08% lower MAE than state-of-the-art methods. Meanwhile, real-time performance analysis confirms the model’s operational viability as it maintains excellent prediction accuracy (RMSE < 15) even at high data arrival rates (100 samples/second) while ensuring processing latency below critical thresholds (10 ms) under typical system loads. Full article