Hui Chen

Underwater target detection faces challenges due to poor image quality, multi-scale variations, and occlusion. This study introduces a novel model integrating multi-scale attention and adaptive feature fusion to enhance detection accuracy. Leveraging the DEIM framework, we propose a Multi-Scale Attention Block (MSAB) for improved feature extraction across scales, a Lightweight Sparse Self-Attention Block (LSSA) for noise suppression, an Adaptive Weighted Downsampling Block (AWDS) for information preservation, and a Context-Guided Feature Fusion Module (CGFM) for intelligent feature integration. Experimental results on URPC and DUO datasets demonstrate significant improvements in AP, AP50, and AP75 metrics compared to the baseline model DEIM, with gains of 3.2%, 3.3%, and 3.4% on URPC, and 3.9%, 3.6%, and 4.7% on DUO, respectively. These findings underscore the effectiveness of our approach in addressing the complexities of underwater environments. https://github.com/EUOD/MSAuod.

With the swift progress of deep learning and its wide application in semantic segmentation, the effect of semantic segmentation has been significantly improved. However, how to achieve a reasonable compromise between accuracy, model size, and inference speed is crucial. In this paper, we propose a lightweight multi-scale asymmetric encoder–decoder network (LMANet) that is designed on the basis of an encoder–decoder structure. First, an optimized bottleneck module is used to extract features from different levels, and different receptive fields are applied to obtain effective information on different scales. Then, a channel-attention module and a feature-extraction module are introduced to constitute the residual structure, and different feature maps are connected by a feature-fusion module to effectively improve segmentation accuracy. Finally, a lightweight multi-scale decoder is designed to recover the image, and a spatial attention module is added to recover the spatial details effectively. This paper has verified the proposed method on the Cityscapes dataset and CamVid dataset and achieved mean intersection over union (mIoU) of 73.9% and 71.3% with the inference speeds of 111 FPS and 118 FPS, respectively, and the number of parameters is only 0.85 M.

To address the problems of low monitoring area coverage rate and the long moving distance of nodes in the process of coverage optimization in wireless sensor networks (WSNs), a multi-strategy improved sparrow search algorithm for coverage optimization in a WSN (IM-DTSSA) is proposed. Firstly, Delaunay triangulation is used to locate the uncovered areas in the network and optimize the initial population of the IM-DTSSA algorithm, which can improve the convergence speed and search accuracy of the algorithm. Secondly, the quality and quantity of the explorer population in the sparrow search algorithm are optimized by the non-dominated sorting algorithm, which can improve the global search capability of the algorithm. Finally, a two-sample learning strategy is used to improve the follower position update formula and to improve the ability of the algorithm to jump out of the local optimum. Simulation results show that the coverage rate of the IM-DTSSA algorithm is increased by 6.74%, 5.04% and 3.42% compared to the three other algorithms. The average moving distance of nodes is reduced by 7.93 m, 3.97 m, and 3.09 m, respectively. The results mean that the IM-DTSSA algorithm can effectively balance the coverage rate of the target area and the moving distance of nodes.

In recent years, chest X-ray (CXR) imaging has become one of the significant tools to assist in the diagnosis and treatment of novel coronavirus pneumonia. However, CXR images have complex-shaped and changing lesion areas, which makes it difficult to identify novel coronavirus pneumonia from the images. To address this problem, a new deep learning network model (BoT-ViTNet) for automatic classification is designed in this study, which is constructed on the basis of ResNet50. First, we introduce multi-headed self-attention (MSA) to the last Bottleneck block of the first three stages in the ResNet50 to enhance the ability to model global information. Then, to further enhance the feature expression performance and the correlation between features, the TRT-ViT blocks, consisting of Transformer and Bottleneck, are used in the final stage of ResNet50, which improves the recognition of complex lesion regions in CXR images. Finally, the extracted features are delivered to the global average pooling layer for global spatial information integration in a concatenated way and used for classification. Experiments conducted on the COVID-19 Radiography database show that the classification accuracy, precision, sensitivity, specificity, and F1-score of the BoT-ViTNet model is 98.91%, 97.80%, 98.76%, 99.13%, and 98.27%, respectively, which outperforms other classification models. The experimental results show that our model can classify CXR images better.