Search Results (1,021)

Background: Accurate segmentation of the left ventricle in echocardiograms is crucial for the diagnosis and monitoring of cardiovascular diseases. However, this process is hindered by the limited availability of high-quality annotated datasets and the inherent complexities of echocardiogram images. Traditional methods often struggle to generalize across varying image qualities and conditions, necessitating a more robust solution. Objectives: This study aims to enhance left ventricular segmentation in echocardiograms by developing a framework that integrates Generative Adversarial Networks (GANs) for synthetic data augmentation with a MultiResUNet architecture, providing a more accurate and reliable segmentation method. Methods: We propose a GAN-based framework that generates synthetic echocardiogram images and their corresponding segmentation masks, augmenting the available training data. The synthetic data, along with real echocardiograms from the EchoNet-Dynamic dataset, were used to train the MultiResUNet architecture. MultiResUNet incorporates multi-resolution blocks, residual connections, and attention mechanisms to effectively capture fine details at multiple scales. Additional enhancements include atrous spatial pyramid pooling (ASPP) and scaled exponential linear units (SELUs) to further improve segmentation accuracy. Results: The proposed approach significantly outperforms existing methods, achieving a Dice Similarity Coefficient of 95.68% and an Intersection over Union (IoU) of 91.62%. This represents improvements of 2.58% in Dice and 4.84% in IoU over previous segmentation techniques, demonstrating the effectiveness of GAN-based augmentation in overcoming data scarcity and improving segmentation performance. Conclusions: The integration of GAN-generated synthetic data and the MultiResUNet architecture provides a robust and accurate solution for left ventricular segmentation in echocardiograms. This approach has the potential to enhance clinical decision-making in cardiovascular medicine by improving the accuracy of automated diagnostic tools, even in the presence of limited and complex training data. Full article

The lack of extensive, varied, and thoroughly annotated datasets impedes the advancement of artificial intelligence (AI) for medical applications, especially colorectal cancer detection. Models trained with limited diversity often display biases, especially when utilized on disadvantaged groups. Generative models (e.g., DALL-E 2, Vector-Quantized Generative Adversarial Network (VQ-GAN)) have been used to generate images but not colonoscopy data for intelligent data augmentation. This study developed an effective method for producing synthetic colonoscopy image data, which can be used to train advanced medical diagnostic models for robust colorectal cancer detection and treatment. Text-to-image synthesis was performed using fine-tuned Visual Large Language Models (LLMs). Stable Diffusion and DreamBooth Low-Rank Adaptation produce images that look authentic, with an average Inception score of 2.36 across three datasets. The validation accuracy of various classification models Big Transfer (BiT), Fixed Resolution Residual Next Generation Network (FixResNeXt), and Efficient Neural Network (EfficientNet) were 92%, 91%, and 86%, respectively. Vision Transformer (ViT) and Data-Efficient Image Transformers (DeiT) had an accuracy rate of 93%. Secondly, for the segmentation of polyps, the ground truth masks are generated using Segment Anything Model (SAM). Then, five segmentation models (U-Net, Pyramid Scene Parsing Network (PSNet), Feature Pyramid Network (FPN), Link Network (LinkNet), and Multi-scale Attention Network (MANet)) were adopted. FPN produced excellent results, with an Intersection Over Union (IoU) of $0.64$, an F1 score of $0.78$, a recall of $0.75$, and a Dice coefficient of 0.77. This demonstrates strong performance in terms of both segmentation accuracy and overlap metrics, with particularly robust results in balanced detection capability as shown by the high F1 score and Dice coefficient. This highlights how AI-generated medical images can improve colonoscopy analysis, which is critical for early colorectal cancer detection. Full article

In this research, a multi-channel target speech enhancement scheme is proposed that is based on deep learning (DL) architecture and assisted by multi-source tracking using a labeled random finite set (RFS) framework. A neural network based on minimum variance distortionless response (MVDR) beamformer is considered as the beamformer of choice, where a residual dense convolutional graph-U-Net is applied in a generative adversarial network (GAN) setting to model the beamformer for target speech enhancement under reverberant conditions involving multiple moving speech sources. The input dataset for this neural architecture is constructed by applying multi-source tracking using multi-sensor generalized labeled multi-Bernoulli (MS-GLMB) filtering, which belongs to the labeled RFS framework, to obtain estimations of the sources’ positions and the associated labels (corresponding to each source) at each time frame with high accuracy under the effect of undesirable factors like reverberation and background noise. The tracked sources’ positions and associated labels help to correctly discriminate the target source from the interferers across all time frames and generate time–frequency (T-F) masks corresponding to the target source from the output of a time-varying, minimum variance distortionless response (MVDR) beamformer. These T-F masks constitute the target label set used to train the proposed deep neural architecture to perform target speech enhancement. The exploitation of MS-GLMB filtering and a time-varying MVDR beamformer help in providing the spatial information of the sources, in addition to the spectral information, within the neural speech enhancement framework during the training phase. Moreover, the application of the GAN framework takes advantage of adversarial optimization as an alternative to maximum likelihood (ML)-based frameworks, which further boosts the performance of target speech enhancement under reverberant conditions. The computer simulations demonstrate that the proposed approach leads to better target speech enhancement performance compared with existing state-of-the-art DL-based methodologies which do not incorporate the labeled RFS-based approach, something which is evident from the 75% ESTOI and PESQ of 2.70 achieved by the proposed approach as compared with the 46.74% ESTOI and PESQ of 1.84 achieved by Mask-MVDR with self-attention mechanism at a reverberation time (RT60) of 550 ms. Full article

High-resolution remote sensing imagery (HRRSI) presents significant challenges for building extraction tasks due to its complex terrain structures, multi-scale features, and rich spectral and geometric information. Traditional methods often face limitations in effectively integrating multi-scale features while maintaining a balance between detailed and global semantic information. To address these challenges, this paper proposes an innovative deep learning network, Multi-Source Multi-Scale Residual Attention Network (MMRAD-Net). This model is built upon the classical encoder–decoder framework and introduces two key components: the GCN OA-SWinT Dense Module (GSTDM) and the Res DualAttention Dense Fusion Block (R-DDFB). Additionally, it incorporates Digital Surface Model (DSM) data, presenting a novel feature extraction and fusion strategy. Specifically, the model enhances building extraction accuracy and robustness through hierarchical feature modeling and a refined cross-scale fusion mechanism, while effectively preserving both detail information and global semantic relationships. Furthermore, we propose a Hybrid Loss, which combines Binary Cross-Entropy Loss (BCE Loss), Dice Loss, and an edge-sensitive term to further improve the precision of building edges and foreground reconstruction capabilities. Experiments conducted on the GF-7 and WHU datasets validate the performance of MMRAD-Net, demonstrating its superiority over traditional methods in boundary handling, detail recovery, and adaptability to complex scenes. On the GF-7 Dataset, MMRAD-Net achieved an F1-score of 91.12% and an IoU of 83.01%. On the WHU Building Dataset, the F1-score and IoU were 94.04% and 88.99%, respectively. Ablation studies and transfer learning experiments further confirm the rationality of the model design and its strong generalization ability. These results highlight that innovations in multi-source data fusion, multi-scale feature modeling, and detailed feature fusion mechanisms have enhanced the accuracy and robustness of building extraction. Full article

Nucleus accurate segmentation is a crucial task in biomedical image analysis. While convolutional neural networks (CNNs) have achieved notable progress in this field, challenges remain due to the complexity and heterogeneity of cell images, especially in overlapping regions of nuclei. To address the limitations of current methods, we propose a mechanism of multiple differential convolution and local-variation attention in CNNs, leading to the so-called multiple differential convolution and local-variation attention U-Net (MDLA-UNet). The multiple differential convolution employs multiple differential operators to capture gradient and direction information, improving the network’s capability to detect edges. The local-variation attention utilizes Haar discrete wavelet transforms for level-1 decomposition to obtain approximate features, and then derives high-frequency features to enhance the global context and local detail variation of the feature maps. The results on the MoNuSeg, TNBC, and CryoNuSeg datasets demonstrated superior segmentation performance of the proposed method for cells having complex boundaries and details with respect to existing methods. The proposed MDLA-UNet presents the ability of capturing fine edges and details in feature maps and thus improves the segmentation of nuclei with blurred boundaries and overlapping regions. Full article

The accurate detection and monitoring of levee cracks is critical for maintaining the structural integrity and safety of flood protection infrastructure. Yet at present the application of using UAV to achieve an automatic, rapid detection of levee cracks is still limited and there is a lack of effective deblurring methods specifically tailored for UAV-based levee crack images. In this study, we present E2S, a novel two-stage framework specifically designed for UAV-based levee crack segmentation, which leverages an unsupervised deblurring technique to enhance image quality. In the first stage, we introduce an Improved CycleGAN model that mainly performs motion deblurring on UAV-captured images, effectively enhancing crack visibility and preserving crucial structural details. The enhanced images are then fed into the second stage, where an Attention U-Net is employed for precise crack segmentation. The experimental results demonstrate that the E2S framework significantly outperforms traditional supervised models, achieving an F1-score of 81.3% and a crack IoU of 71.84%, surpassing the best-performing baseline, Unet++. The findings confirm that the integration of unsupervised image enhancement can substantially benefit downstream segmentation tasks, providing a robust and scalable solution for automated levee crack monitoring. Full article

The state monitoring of power load equipment plays a crucial role in ensuring its normal operation. However, in densely deployed environments, the target equipment often exhibits low clarity, making real-time warnings challenging. In this study, a load equipment segmentation and assessment method based on multi-source tensor feature fusion (LSA-MT) is proposed. First, a lightweight residual block based on the attention mechanism is introduced into the backbone network to emphasize key features of load devices and enhance target segmentation efficiency. Second, a 3D edge detail feature perception module is designed to facilitate multi-scale feature fusion while preserving boundary detail features of different devices, thereby improving local recognition accuracy. Finally, tensor decomposition and reorganization are employed to guide visual feature reconstruction in conjunction with equipment monitoring images, while tensor mapping of equipment monitoring data is utilized for automated fault classification. The experimental results demonstrate that LSE-MT produces visually clearer segmentations compared to models such as the classic UNet++ and the more recent EGE-UNet when segmenting multiple load devices, achieving Dice and mIoU scores of 92.48 and 92.90, respectively. Regarding classification across the four datasets, the average accuracy can reach 92.92%. These findings fully demonstrate the effectiveness of the LSA-MT method in load equipment fault alarms and grid operation and maintenance. Full article

Background: Infant brain tissue segmentation from MRI data is a critical task in medical imaging, particularly challenging due to the evolving nature of tissue contrasts in the early months of life. The difficulty increases as gray matter (GM) and white matter (WM) intensities converge, making accurate segmentation challenging. This study aims to develop an improved U-net-based model to enhance the precision of automatic segmentation of cerebro-spinal fluid (CSF), GM, and WM in 10 infant brain MRIs using the iSeg-2017 dataset. Methods: The proposed method utilizes a U-net architecture with (2+1)Dconvolutional layers and skip connections. Preprocessing includes intensity normalization using histogram alignment to standardize MRI data across different records. The model was trained on the iSeg-2017 dataset, which comprises T1-weighted and T2-weighted MRI data from ten infant subjects. Cross-validation was performed to evaluate the model’s segmentation performance. Results: The model achieved an average accuracy of 92.2%, improving on previous methods by 0.7%. Sensitivity, precision, and Dice similarity scores were used to evaluate the performance, showing high levels of accuracy across different tissue types. The model demonstrated a slight bias toward misclassifying GM and WM, indicating areas for potential improvement. Conclusions: The results suggest that the U-net architecture is highly effective in segmenting infant brain tissues from MRI data. Future work will explore enhancements such as attention mechanisms and dual-network processing for further improving segmentation accuracy. Full article

Aimed at the limitation that existing hyperspectral classification methods were mainly oriented to small-scale images, this paper proposed a new large-scale hyperspectral remote sensing image classification method, LS3EU-Net++ (Lightweight Encoder and Integrated Spatial Spectral Squeeze and Excitation U-Net++). The method optimized the U-Net++ architecture by introducing a lightweight encoder and combining the Spatial Spectral Squeeze and Excitation (S3E) Attention Module, which maintained the powerful feature extraction capability while significantly reducing the training cost. In addition, the model employed a composite loss function combining focal loss and Jaccard loss, which could focus more on difficult samples, thus improving pixel-level accuracy and classification results. To solve the sample imbalance problem in hyperspectral images, this paper also proposed a data enhancement strategy based on “copy–paste”, which effectively increased the diversity of the training dataset. Experiments on large-scale satellite hyperspectral remote sensing images from the Zhuhai-1 satellite demonstrated that LS3EU-Net++ exhibited superiority over the U-Net++ benchmark. Specifically, the overall accuracy (OA) was improved by 5.35%, and the mean Intersection over Union (mIoU) by 12.4%. These findings suggested that the proposed method provided a robust solution for large-scale hyperspectral image classification, effectively balancing accuracy and computational efficiency. Full article

Prostate cancer (PCa) is one of the most common tumors diagnosed in men worldwide, with approximately 1.7 million new cases expected by 2030. Most cancerous lesions in PCa are located in the peripheral zone (PZ); therefore, accurate identification of the location of the lesion is essential for effective diagnosis and treatment. Zonal segmentation in magnetic resonance imaging (MRI) scans is critical and plays a key role in pinpointing cancerous regions and treatment strategies. In this work, we report on the development of three advanced neural network-based models: one based on ensemble learning, one on Meta-Net, and one on YOLO-V8. They were tailored for the segmentation of the central gland (CG) and PZ using a small dataset of 90 MRI scans for training, 25 MRIs for validation, and 24 scans for testing. The ensemble learning method, combining U-Net-based models (Attention-Res-U-Net, Vanilla-Net, and V-Net), achieved an IoU of 79.3% and DSC of 88.4% for CG and an IoU of 54.5% and DSC of 70.5% for PZ on the test set. Meta-Net, used for the first time in segmentation, demonstrated an IoU of 78% and DSC of 88% for CG, while YOLO-V8 outperformed both models with an IoU of 80% and DSC of 89% for CG and an IoU of 58% and DSC of 73% for PZ. Full article

The Qinghai–Tibet Plateau (QTP), one of China’s most snow-rich regions, has an extremely fragile ecosystem, with drought being the primary driver of ecological degradation. Given that the water resources in this region predominantly exist in the form of snow, high-spatiotemporal-resolution snow mapping is essential for understanding snow distribution and managing snow water resources effectively. However, although FY-4A/AGRI is capable of obtaining wide-area remote sensing data, only the first to third bands have a resolution of 1 km, which greatly limits its ability to produce high-resolution snow maps. This study proposes U-SwinFusionNet (USFNet), a deep learning-based snow cover retrieval algorithm that leverages the multi-scale advantages of FY-4A/AGRI remote sensing data in the shortwave infrared and visible bands. By integrating 1 km and 2 km resolution remote sensing imagery with auxiliary terrain information, USFNet effectively enhances snow cover mapping accuracy. The proposed model innovatively combines Swin Transformer and convolutional neural networks (CNNs) to capture both global contextual information and local spatial details. Additionally, an Attention Feature Fusion Module (AFFM) is introduced to align and integrate features from different modalities through an efficient attention mechanism, while the Feature Complementation Module (FCM) facilitates interactions between the encoded and decoded features. As a result, USFNet produces snow cover maps with a spatial resolution of 1 km. Experimental comparisons with Artificial Neural Networks (ANNs), Random Forest (RF), U-Net, and ResNet-FSC demonstrate that USFNet exhibits superior robustness, enhanced snow cover continuity, and lower error rates. The model achieves a correlation coefficient of 0.9126 and an R² of 0.7072. Compared to the MOD10A1 snow product, USFNet demonstrates an improved sensitivity to fragmented and low-snow-cover areas while ensuring more natural snow boundary transitions. Full article

The solidified state after the melting of the forming layer in the laser powder bed fusion (LPBF) directly reflects the final forming quality. Compared with the powder layer and the melt pool, it is easier to recognize and remove the defects of contour parts in time by monitoring and processing the solidified region after the melting of the forming layer. To explore the application of a solidified region image in defect contour detection of the forming layer, an improved image segmentation model based on TranUNet is designed to extract the image features of the solidified region as process data, on which basis this paper analyzes the similarities and differences between forming process data and CT scanning results data. Addressing the characteristics of large data volume and significant feature scale variation in the solidified region image obtained during the LPBF process, an SA-TransUNet semantic segmentation model integrating SE attention mechanism and ASPP multi-scale feature extraction module is developed to achieve high-precision solidified region image segmentation, with an $IoU$ and a dice coefficient index up to 94.24% and 97.02%, respectively. By extracting the solidified region image of the LPBF forming layer through this model and calculating the geometric feature values of its contour, a comparative analysis is conducted with the corresponding contour geometric feature values of the formed part CT scan image, which verifies the feasibility of the solidified region image extraction method proposed in this paper for contour defect detection. Full article

Uncooled infrared imaging systems have significant potential in industrial hazardous gas leak detection. However, the use of narrowband filters to match gas spectral absorption peaks leads to a low level of incident energy captured by uncooled infrared cameras. This results in a mixture of fixed pattern noise and Gaussian noise, while existing denoising methods for uncooled infrared images struggle to effectively address this mixed noise, severely hindering the extraction and identification of actual gas leak plumes. This paper presents a UNet-structured dual-encoder denoising network specifically designed for narrowband uncooled infrared images. Based on the distinct characteristics of Gaussian random noise and row–column stripe noise, we developed a basic scale residual attention (BSRA) encoder and an enlarged scale residual attention (ESRA) encoder. These two encoder branches perform noise perception and encoding across different receptive fields, allowing for the fusion of noise features from both scales. The combined features are then input into the decoder for reconstruction, resulting in high-quality infrared images. Experimental results demonstrate that our method effectively denoises composite noise, achieving the best results according to both objective metrics and subjective evaluations. This research method significantly enhances the signal-to-noise ratio of narrowband uncooled infrared images, demonstrating substantial application potential in fields such as industrial hazardous gas detection, remote sensing imaging, and medical imaging. Full article

The primary objective of this research was to develop an efficient method for accurately identifying and localizing weeds in ecological tea garden environments, aiming to enhance the quality and yield of tea production. Weed competition poses a significant challenge to tea production, particularly due to the small size of weed plants, their color similarity to tea trees, and the complexity of their growth environment. A dataset comprising 5366 high-definition images of weeds in tea gardens has been compiled to address this challenge. An enhanced U-Net model, incorporating a Double Attention Mechanism and an Atrous Spatial Pyramid Pooling module, is proposed for weed recognition. The results of the ablation experiments show that the model significantly improves the recognition accuracy and the Mean Intersection over Union (MIoU), which are enhanced by 4.08% and 5.22%, respectively. In addition, to meet the demand for precise weed management, a method for determining the center of weed plants by integrating the center of mass and skeleton structure has been developed. The skeleton was extracted through a preprocessing step and a refinement algorithm, and the relative positional relationship between the intersection point of the skeleton and the center of mass was cleverly utilized to achieve up to 82% localization accuracy. These results provide technical support for the research and development of intelligent weeding equipment for tea gardens, which helps to maintain the ecology of tea gardens and improve production efficiency and also provides a reference for weed management in other natural ecological environments. Full article

Gamma-ray spectroscopy is essential in nuclear science, enabling the identification of radioactive materials through energy spectrum analysis. However, high count rates lead to pile-up effects, resulting in spectral distortions that hinder accurate isotope identification and activity estimation. This phenomenon highlights the need for automated and precise approaches to pile-up correction. We propose a novel deep learning (DL) framework plugging count rate information of pile-up signals with a 2D attention U-Net for energy spectrum recovery. The input to the model is an Energy–Duration matrix constructed from preprocessed pulse signals. Temporal and spatial features are jointly extracted, with count rate information embedded to enhance robustness under high count rate conditions. Training data were generated using an open-source simulator based on a public gamma spectrum database. The model’s performance was evaluated using Kullback–Leibler (KL) divergence, Mean Squared Error (MSE) Energy Resolution (ER), and Full Width at Half Maximum (FWHM). Results indicate that the proposed framework effectively predicts accurate spectra, minimizing errors even under severe pile-up effects. This work provides a robust framework for addressing pile-up effects in gamma-ray spectroscopy, presenting a practical solution for automated, high-accuracy spectrum estimation. The integration of temporal and spatial learning techniques offers promising prospects for advancing high-activity nuclear analysis applications. Full article