Deep Learning Model for Precipitation Nowcasting Based on Residual and Attention Mechanisms

Air pollution, particularly fine particulate matter (PM2.5), poses significant public health and environmental risks. This study explores the effectiveness of spatiotemporal graph neural networks (ST-GNNs) in forecasting PM2.5 concentrations by integrating remote-sensing hyperspectral indices with traditional meteorological and pollutant data. The model was evaluated using data from Switzerland and the Gauteng province in South Africa, with datasets spanning from January 2016 to December 2021. Key performance metrics, including root mean squared error (RMSE), mean absolute error (MAE), probability of detection (POD), critical success index (CSI), and false alarm rate (FAR), were employed to assess model accuracy. For Switzerland, the integration of spectral indices improved RMSE from 1.4660 to 1.4591, MAE from 1.1147 to 1.1053, CSI from 0.8345 to 0.8387, POD from 0.8961 to 0.8972, and reduced FAR from 0.0760 to 0.0719. In Gauteng, RMSE decreased from 6.3486 to 6.2319, MAE from 4.4891 to 4.4066, CSI from 0.9555 to 0.9560, and POD from 0.9699 to 0.9732, while FAR slightly increased from 0.0154 to 0.0181. Error analysis revealed that while the initial one-day ahead forecast without spectral indices had a marginally lower error, the dataset with spectral indices outperformed from the two-day ahead mark onwards. The error for Swiss monitoring stations stabilized over longer prediction lengths, indicating the robustness of the spectral indices for extended forecasts. The study faced limitations, including the exclusion of the Planetary Boundary Layer (PBL) height and K-index, lack of terrain data for South Africa, and significant missing data in remote sensing indices. Despite these challenges, the results demonstrate that ST-GNNs, enhanced with hyperspectral data, provide a more accurate and reliable tool for PM2.5 forecasting. Future work will focus on expanding the dataset to include additional regions and further refining the model by incorporating additional environmental variables. This approach holds promise for improving air quality management and mitigating health risks associated with air pollution.

BackgroundGrain count is crucial to wheat yield composition and estimating yield parameters. However, traditional manual counting methods are time-consuming and labor-intensive. This study developed an advanced deep learning technique for the segmentation counting model of wheat grains. This model has been rigorously tested on three distinct wheat varieties: ‘Bainong 307’, ‘Xinmai 26’, and ‘Jimai 336’, and it has achieved unprecedented predictive counting accuracy.MethodThe images of wheat ears were taken with a smartphone at the late stage of wheat grain filling. We used image processing technology to preprocess and normalize the images to 480*480 pixels. A CBAM-HRNet wheat grain segmentation counting deep learning model based on the Convolutional Block Attention Module (CBAM) was constructed by combining deep learning, migration learning, and attention mechanism. Image processing algorithms and wheat grain texture features were used to build a grain counting and predictive counting model for wheat grains.ResultsThe CBAM-HRNet model using the CBAM was the best for wheat grain segmentation. Its segmentation accuracy of 92.04%, the mean Intersection over Union (mIoU) of 85.21%, the category mean pixel accuracy (mPA) of 91.16%, and the recall rate of 91.16% demonstrate superior robustness compared to other models such as HRNet, PSPNet, DeeplabV3+ , and U-Net. Method I for spike count, which calculates twice the number of grains on one side of the spike to determine the total number of grains, demonstrates a coefficient of determination R2 of 0.85, a mean absolute error (MAE) of 1.53, and a mean relative error (MRE) of 2.91. In contrast, Method II for spike count involves summing the number of grains on both sides to determine the total number of grains, demonstrating a coefficient of determination R2 of 0.92, an MAE) of 1.15, and an MRE) of 2.09%.ConclusionsImage segmentation algorithm of the CBAM-HRNet wheat spike grain is a powerful solution that uses the CBAM to segment wheat spike grains and obtain richer semantic information. This model can effectively address the challenges of small target image segmentation and under-fitting problems in training. Additionally, the spike grain counting model can quickly and accurately predict the grain count of wheat, providing algorithmic support for efficient and intelligent wheat yield estimation.

MRI image classification of brain tumors is critical for accurate and early diagnosis. New developments in deep learning have revealed that inserting attention mechanisms into convolutional neural networks can greatly improve classification performance. This work assesses the effectiveness of Squeeze-and-Excitation (SE) and Convolutional Block Attention Module (CBAM) integrated with the ResNet50 model for brain tumor classification. The SE module recalibrates channel-wise feature responses, while CBAM enhances feature representation by combining channel and spatial attention. The SE and CBAM are combined with ResNet50, greatly increasing classification accuracy, precision, and recall when compared to the basic model, according to experimental results on MRI datasets. The suggested model employs attention mechanisms to focus valuable information selectively and suppress irrelevant information. The experiments on publicly available brain tumor datasets display great improvements over basic CNN models, with precision, recall, accuracy, as well as F1 score at 0.9914, 0.9903, 0.9945, and 0.9908, respectively. From these results, the importance of attention mechanisms in deep learning models for medical imaging is highlighted, which suggests that SE and CBAM modules can be available as more dependable and effective instruments for brain tumor classification in clinical settings. Future studies should investigate transformer-based and hybrid attention techniques to enhance automated brain tumor categorization.

Precipitation nowcasting play a vital role in preventing meteorological disasters and doppler radar data acts as an important input for nowcasting models. The traditional extrapolation method is difficult to model highly nonlinear echo movements. The key challenge of the nowcasting mission lies in achieving high-precision radar echo extrapolation. In recent years, machine learning has made a great progress in the extrapolation of weather radar echoes. However, most of models neglect the multi-modal characteristics of radar echo data, resulting in blurred and unrealistic prediction images. This paper aims to solve this problem by utilizing the feature of the GAN that can enhance the multi-modal distribution modelling, and design the radar echo extrapolation model of GAN-argcPredNet. The model composed of argcPredNet generator and a convolutional neural network discriminator. In generator, a gate control data memory and output are designed in the rgcLSTM prediction unit of the generator, thereby reducing the loss of spatiotemporal information. In discriminator, model uses a dual-channel input method, which enables it to strictly score according to the true echo distribution, and has a more powerful discrimination ability. Through experiments on the radar data set of Shenzhen, China, the results show that the radar echo hit rate (POD) and critical success index (CSI) increased by 5.5 % and %10.4 % compared with rgcPredNet, the false alarm rate (FAR) is reduced by 15 %~20 %. From the comparison of the result graph and the evaluation index, we also found a problem. The recursive prediction method will produce the phenomenon that the prediction result will gradually deviate from the true value over time. In addition, the accuracy of high-intensity echo extrapolation is relatively low. This is a question worthy of further investigation. In the future, we will continue to conduct research from these two directions.

This study utilizes the Breast Ultrasound Image (BUSI) dataset to present a deep learning technique for breast tumor segmentation based on a modified UNet architecture. To improve segmentation accuracy, the model integrates attention mechanisms, such as the Convolutional Block Attention Module (CBAM) and Non-Local Attention, with advanced encoder architectures, including ResNet, DenseNet, and EfficientNet. These attention mechanisms enable the model to focus more effectively on relevant tumor areas, resulting in significant performance improvements. Models incorporating attention mechanisms outperformed those without, as reflected in superior evaluation metrics. The effects of Dice Loss and Binary Cross-Entropy (BCE) Loss on the model’s performance were also analyzed. Dice Loss maximized the overlap between predicted and actual segmentation masks, leading to more precise boundary delineation, while BCE Loss achieved higher recall, improving the detection of tumor areas. Grad-CAM visualizations further demonstrated that attention-based models enhanced interpretability by accurately highlighting tumor areas. The findings denote that combining advanced encoder architectures, attention mechanisms, and the UNet framework can yield more reliable and accurate breast tumor segmentation. Future research will explore the use of multi-modal imaging, real-time deployment for clinical applications, and more advanced attention mechanisms to further improve segmentation performance.

Abstract. This study explores the enhancement of UNet-based semantic segmentation for photovoltaic (PV) panels in remote sensing images by integrating attention mechanisms. Given the critical role of solar energy in achieving global sustainability, accurate PV panel detection is essential for effective energy management. Using the high-resolution PV01 dataset, which includes UAV-captured rooftop PV samples, we evaluate the impact of four distinct attention modules: Convolutional Block Attention Module (CBAM), Squeeze-and-Excitation Networks (SE-Net), Efficient Channel Attention (ECA-Net), and Coordinate Attention (CA) on segmentation performance. Comparative analysis demonstrates that UNet models with SE and CA modules substantially outperform the baseline, achieving the highest scores in Average Accuracy (AA), Average Precision (AP), Average Recall (AR), mean Intersection over Union (mIoU), and Average F1-score (AF). In particular, UNet + SE achieved an AA of 0.9809, AP of 0.9756, AR of 0.9629, AF of 0.9692, and mIoU of 0.9403, highlighting the efficacy of attention mechanisms in refining feature representations and advancing PV panel segmentation, thereby contributing to large-scale solar energy monitoring and deployment.

To reduce the chance of being infected by the COVID-19, wearing masks correctly when entering and leaving public places has become the most feasible and effective ways to prevent the spread of the virus. It is a concern to how to quickly and accurately detect whether a face is worn a mask correctly while reduce missed detection and false detection in practical applied scenarios. In this paper, an improved algorithm is proposed based on the YOLO-v4 algorithm. The attention mechanism module is added to the appropriate network level to enhance the key feature points of face wearing masks and suppress useless information. Apart from that, three attention mechanism modules are added to different layers of the YOLO-v4 network for ablation experiments, including CBAM (convolutional block attention module), SENet (squeeze-and-excitation networks) and CANet (coordinate attention networks). The path-aggregation network and feature pyramid are used to extract features from images. Two network models were compared and improved in the experiment, and it is found that adding the dual-channel attention mechanism CBAM before the three YOLO heads of YOLOv4 and in the neck network had better detection performance than the single channel attention mechanism SENet and the coordinated attention mechanism CANet. The experimental results show that when the attention module CBAM and the YOLO-v4 model are integrated, the accuracy of the selected MAFA + WIDER Face dataset reaches the highest value of 93.56%, which is 4.66% higher than that of the original YOLO-v4.

Brain-computer interfaces (BCI) allow for the brain to communicate with electronic devices. Concerning the BCI paradigms, motor imagery uses brain signals to decode an imagined movement. However, this is a hard task given the low signal-to-noise ratio. Usually, the main steps in BCI models are pre-processing, feature extraction, and classification. In recent years, Convolutional Neural Networks (CNNs) have been gaining relevance in several areas of science due to their feature extraction, translation invariance, and parameter sharing capa-bilities. Another, more recent way of feature extraction is using attention mechanisms, which are layers of neural networks based on human attention and have the ability to highlight important features. A variation of the attention mechanism is the Convolutional Block Attention Module, which combines the CNN structure with the attention mechanism. In this work, we propose a new model that joins the core architecture of EEGNet, a compact CNN widely used in the literature, with the Convolutional Block Attention Module and residual connections. The residual connections were introduced to lower data degradation throughout the model. The results highlight the residual connection's importance for the performance of the model. The proposed model obtained a kappa result 5.2% better than the EEGNet with a p-value less than 0.01 on BCI Competition IV dataset 2a, which is a well-known dataset for Motor Imagery. Furthermore, the proposal was better than EEGNet for most subjects and had the best-worst case.

Weather radar echo prediction method based on convolution neural network and Long Short-Term memory networks for sustainable e-agriculture

As coal mining deepens, ground fissures in mining areas pose significant risks to safety and the environment. Traditional geological exploration methods are inefficient and costly, making the precise detection of large-scale fissures difficult. Deep learning methods using unmanned aerial vehicle (UAV) images have become a popular approach for fissure detection, although challenges such as insufficient accuracy and poor adaptability in complex backgrounds remain. This study proposes an residual-depthwise separable convolution UNet (RDC-UNet) model to address these issues, building on U-Net by incorporating residual connections (RC), depthwise separable convolutions (DSC), and the convolutional block attention module (CBAM) attention mechanism. The model was trained on a dataset of 300 UAV images from a DJI Mavic 3e and outperformed the benchmark models, achieving 86.12 % mPrecision, 75.01 % mRecall, 70.72 % mIoU, and 0.7951 mF1. Ablation experiments show that removing any core module leads to a performance drop, with the DSC module reducing mF1 by 3.28 %, CBAM decreasing mIoU by 2.67 %, and RC lowering mRecall by 3.96 %. RDC-UNet is highly efficient, requiring only 7.8 million parameters, with an inference time of 75 milliseconds and a memory footprint of 180 MB, much lower than other models. This makes it well-suited for real-time UAV-based fissure monitoring. The RDC-UNet model offers high accuracy and efficiency, making it an ideal solution for cost-effective real-time monitoring of ground fissures in mining areas.

Rapid detection and identification of landslide areas are very important for disaster prevention and mitigation. Aiming at the problems of time-consuming and labor-intensive traditional landslide information extraction methods and low recognition efficiency, a remote sensing landslide recognition method based on LinkNet, and convolution attention module was proposed. The model adopts the coding-decoding structure to improve the operation efficiency. The Convolutional Block Attention Module (CBAM) is applied to optimize the weight allocation from both channel and spatial dimensions to highlight the landslide feature information. And compared with the traditional U-Net and LinkNet models. The results show that the CBAM-LinkNet model has excellent performance in remote sensing landslide identification, which provides the possibility for rapid and accurate landslide identification.

Maritime vessel detection in nearshore waters is a fundamental capability for artificial intelligence (AI)-enabled maritime transportation systems, including coastal monitoring, traffic management, and digital maritime services. Although attention mechanisms are widely incorporated into YOLO-based detectors, their relative effectiveness in marine environments under strictly controlled experimental conditions remains insufficiently clarified. This study presents a systematic comparison of Coordinate Attention (CA), Convolutional Block Attention Module (CBAM), and CLIP-based semantic fusion within a unified YOLOv8n framework for binary discrimination between ships and fishing boats in cluttered coastal imagery. All model variants were trained under identical data partitions and optimization settings to isolate architectural effects. The experimental results show that CA achieves the highest localization robustness (mAP@0.5:0.95 = 0.6127) and substantially improves precision (+7.13% over baseline), while CBAM provides the most balanced performance with the highest F1-score. In contrast, CLIP-based semantic fusion consistently degrades detection reliability, indicating limitations of global vision–language representations in small-scale maritime datasets. Precision–Recall and F1 analyses further reveal architecture-specific confidence calibration behaviors relevant to deployment-sensitive maritime applications. The findings provide practical guidance for selecting attention mechanisms in AI-driven maritime perception systems and support reliable AI integration in marine science and engineering applications.

Background: The accurate classification between malignant and benign breast lesions detected on mammograms is a crucial but difficult challenge for reducing false-positive recall rates and improving the efficacy of breast cancer screening. Objective: This study aims to optimize a new deep transfer learning model by implementing a novel attention mechanism in order to improve the accuracy of breast lesion classification. Methods: ResNet50 is selected as the base model to develop a new deep transfer learning model. To enhance the accuracy of breast lesion classification, we propose adding a convolutional block attention module (CBAM) to the standard ResNet50 model and optimizing a new model for this task. We assembled a large dataset with 4280 mammograms depicting suspicious soft-tissue mass-type lesions. A region of interest (ROI) is extracted from each image based on lesion center. Among them, 2480 and 1800 ROIs depict verified benign and malignant lesions, respectively. The image dataset is randomly split into two subsets with a ratio of 9:1 five times to train and test two ResNet50 models with and without using CBAM. Results: Using the area under ROC curve (AUC) as an evaluation index, the new CBAM-based ResNet50 model yields AUC = 0.866 ± 0.015, which is significantly higher than that obtained by the standard ResNet50 model (AUC = 0.772 ± 0.008) (p < 0.01). Conclusion: This study demonstrates that although deep transfer learning technology attracted broad research interest in medical-imaging informatic fields, adding a new attention mechanism to optimize deep transfer learning models for specific application tasks can play an important role in further improving model performances.

The stability of arc bubble is a crucial indicator of underwater wet welding process. However, limited research exists on detecting arc bubble edges in such environments, and traditional algorithms often produce blurry and discontinuous results. To address these challenges, we propose a novel arc bubble edge detection method based on deep transfer learning for processing underwater wet welding images. The proposed method integrates two training stages: pre-training and fine-tuning. In the pre-training stage, a large source domain dataset is used to train VGG16 as a feature extractor. In the fine-tuning stage, we introduce the Attention-Scale-Semantics (ASS) model, which consists of a Convolutional Block Attention Module (CBAM), a Scale Fusion Module (SCM) and a Semantic Fusion Module (SEM). The ASS model is further trained on a small target domain dataset specific to underwater wet welding to fine-tune the model parameters. The CBAM can adaptively weight the feature maps, focusing on more crucial features to better capture edge information. The SCM training method maximizes feature utilization and simplifies training by combining multi-scale features. Additionally, the skip structure of SEM effectively mitigates semantic loss in the high-level network, enhancing the accuracy of edge detection. On the BSDS500 dataset and a self-constructed underwater wet welding dataset, the ASS model was evaluated against conventional edge detection models—Richer Convolutional Features (RCF), Fully Convolutional Network (FCN), and UNet—as well as state-of-the-art models LDC and TEED. In terms of Mean Absolute Error (MAE), accuracy, and other evaluation metrics, the ASS model consistently outperforms these models, demonstrating edge detection capabilities that are both effective and stable in detecting arc bubble edges in underwater wet welding images.

Wheat head identification and localization are crucial for field management and yield estimation. Traditional wheat head detection relies on labor-intensive manual counting. To address this issue, this study establishes a densely populated wheat head dataset and correspondingly proposes an efficient, rapid real-time model of CB-YOLO, which incorporates a Convolutional Block Attention Module (CBAM) mechanism (a sophisticated fusion of spatial attention mechanism and channel attention mechanism) so as to integrate channel and spatial features. Also, with an incorporated Bidirectional Feature Pyramid Network (BiFPN) to the neck, crowded wheat head detection was enhanced by establishing connections from the backbone to the bottom-up path. This enhancement surpasses the performance of the basic YOLOv5L model, achieving [Formula: see text] of 94.3% and [Formula: see text] of 53.9%. Among 10 classical detection models, CB-YOLO has a relatively small parameter size of 44[Formula: see text]Mb and high computing speed of 108[Formula: see text]G FLOPs. Additionally, ablation experiments demonstrate that compared to YOLOv5L without CBAM and BiFPN, the feature fusion module improves [Formula: see text] from 93.2% to 94.3%, indicating its effectiveness in enhancing detection performance. In conclusion, CB-YOLO exhibits outstanding identification performance, showing its feasibility in practical agricultural applications, and future work will focus on improving its localization ability.