Global Modeling Capability Research Articles

HA-Unet: A Modified Unet Based on Hybrid Attention for Urban Water Extraction in SAR Images

Urban water plays a significant role in the urban ecosystem, but urban water extraction is still a challenging task in automatic interpretation of synthetic aperture radar (SAR) images. The influence of radar shadows and strong scatters in urban areas may lead to misclassification in urban water extraction. Nevertheless, the local features captured by convolutional layers in Convolutional Neural Networks (CNNs) are generally redundant and cannot make effective use of global information to guide the prediction of water pixels. To effectively emphasize the identifiable water characteristics and fully exploit the global information of SAR images, a modified Unet based on hybrid attention mechanism is proposed to improve the performance of urban water extraction in this paper. Considering the feature extraction ability and the global modeling capability in SAR image segmentation, the Channel and Spatial Attention Module (CSAM) and the Multi-head Self-Attention Block (MSAB) are both introduced into the proposed Hybrid Attention Unet (HA-Unet). In this work, Resnet50 is adopted as the backbone of HA-Unet to extract multi-level features of SAR images. During the feature extraction process, CSAM based on local attention is adopted to enhance the meaningful water features and ignore unnecessary features adaptively in feature maps of two shallow layers. In the last two layers of the backbone, MSAB is introduced to capture the global information of SAR images to generate global attention. In addition, two global attention maps generated by MSAB are aggregated together to reconstruct the spatial feature relationship of SAR images from high-resolution feature maps. The experimental results on Sentinel-1A SAR images show that the proposed urban water extraction method has a strong ability to extract water bodies in the complex urban areas. The ablation experiment and visualization results vividly indicate that both CSAM and MSAB contribute significantly to extracting urban water accurately and effectively.

Transformer and group parallel axial attention co-encoder for medical image segmentation

U-Net has become baseline standard in the medical image segmentation tasks, but it has limitations in explicitly modeling long-term dependencies. Transformer has the ability to capture long-term relevance through its internal self-attention. However, Transformer is committed to modeling the correlation of all elements, but its awareness of local foreground information is not significant. Since medical images are often presented as regional blocks, local information is equally important. In this paper, we propose the GPA-TUNet by considering local and global information synthetically. Specifically, we propose a new attention mechanism to highlight local foreground information, called group parallel axial attention (GPA). Furthermore, we effectively combine GPA with Transformer in encoder part of model. It can not only highlight the foreground information of samples, but also reduce the negative influence of background information on the segmentation results. Meanwhile, we introduced the sMLP block to improve the global modeling capability of network. Sparse connectivity and weight sharing are well achieved by applying it. Extensive experiments on public datasets confirm the excellent performance of our proposed GPA-TUNet. In particular, on Synapse and ACDC datasets, mean DSC(%) reached 80.37% and 90.37% respectively, mean HD95(mm) reached 20.55 and 1.23 respectively.

A Novel Method for Ground-Based Cloud Image Classification Using Transformer

In recent years, convolutional neural networks (CNNs) have achieved competitive performance in the field of ground-based cloud image (GCI) classification. Proposed CNN-based methods can fully extract the local features of images. However, due to the locality of the convolution operation, they cannot well establish the long-range dependencies between the images, and thus they cannot extract the global features of images. Transformer has been applied to computer vision with great success due to its powerful global modeling capability. Inspired by it, we propose a Transformer-based GCI classification method that combines the advantages of the CNN and Transformer models. Firstly, the CNN model acts as a low-level feature extraction tool to generate local feature sequences of images. Then, the Transformer model is used to learn the global features of the images by efficiently extracting the long-range dependencies between the sequences. Finally, a linear classifier is used for GCI classification. In addition, we introduce a center loss function to address the problem of the simple cross-entropy loss not adequately supervising feature learning. Our method is evaluated on three commonly used datasets: ASGC, CCSN, and GCD. The experimental results show that the method achieves 94.24%, 92.73%, and 93.57% accuracy, respectively, outperforming other state-of-the-art methods. It proves that Transformer has great potential to be applied to GCI classification tasks.

Building Damage Assessment Based on Siamese Hierarchical Transformer Framework

The rapid and accurate damage assessment of buildings plays a critical role in disaster response. Based on pairs of pre- and post-disaster remote sensing images, effective building damage level assessment can be conducted. However, most existing methods are based on Convolutional Neural Network, which has limited ability to learn the global context. An attention mechanism helps ameliorate this problem. Hierarchical Transformer has powerful potential in the remote sensing field with strong global modeling capability. In this paper, we propose a novel two-stage damage assessment framework called SDAFormer, which embeds a symmetric hierarchical Transformer into a siamese U-Net-like network. In the first stage, the pre-disaster image is fed into a segmentation network for building localization. In the second stage, a two-branch damage classification network is established based on weights shared from the first stage. Then, pre- and post-disaster images are delivered to the network separately for damage assessment. Moreover, a spatial fusion module is designed to improve feature representation capability by building pixel-level correlation, which establishes spatial information in Swin Transformer blocks. The proposed framework achieves significant improvement on the large-scale building damage assessment dataset—xBD.

Cross-Linguistic Similarity Evaluation Techniques Based on Deep Learning

For the cross-linguistic similarity problem, a twin network model with ordered neuron long- and short-term memory neural network as a subnet is proposed. The model fuses bilingual word embeddings and encodes the representation of input sequences by ordered neuron long- and short-term memory neural networks. Based on this, the distributed semantic vector representation of the sentences is jointly constructed by using the global modelling capability of the fully connected network for higher-order semantic extraction. The final output part is the similarity of the bilingual sentences and is optimized by optimizing the parameters of each layer in the framework. Multiple experiments on the dataset show that the model achieves 81.05% accuracy, which effectively improves the accuracy of text similarity and converges faster and improves the semantic text analysis to some extent.

Hybrid Architecture Based on CNN and Transformer for Strip Steel Surface Defect Classification

Strip steel surface defects occur frequently during the manufacturing process, and these defects cause hidden risks in the use of subsequent strip products. Therefore, it is crucial to classify the strip steel’s surface defects accurately and efficiently. Most classification models of strip steel surface defects are generally based on convolutional neural networks (CNNs). However, CNNs, with local receptive fields, do not have admirable global representation ability, resulting in poor classification performance. To this end, we proposed a hybrid network architecture (CNN-T), which merges CNN and Transformer encoder. The CNN-T network has both strong inductive biases (e.g., translation invariance, locality) and global modeling capability. Specifically, CNN first extracts low-level and local features from the images. The Transformer encoder then globally models these features, extracting abstract and high-level semantic information and finally sending them to the multilayer perceptron classifier for classification. Extensive experiments show that the classification performance of CNN-T outperforms pure Transformer networks and CNNs (e.g., GoogLeNet, MobileNet v2, ResNet18) on the NEU-CLS dataset (training ratio is 80%) with a 0.28–2.23% improvement in classification accuracy, with fewer parameters (0.45 M) and floating-point operations (0.12 G).

Extended Matrix Approach for the Identification of Takagi-Sugeno Fuzzy Model

A new, simple and less computational approach is presented to improve the local and global modeling capability of Takagi-Sugeno (T-S) fuzzy model. The main aim is obtaining high function approximation accuracy and fast convergence. The main drawback of T-S identification method is that it can not be applied when the membership functions are overlapped by pairs. This restricts the application of the T-S model because this type of membership function has been widely used during the last two decades in the stability, controller design of fuzzy systems and is popular in industrial control applications. The method developed here can be considered as a generalized version of T-S fuzzy identification method with optimized performance in approximating nonlinear functions. Various examples are chosen to examine the remarkable performance of the proposed method and the high accuracy obtained in approximating nonlinear systems locally and globally in comparison with the original T-S model.

The dynamics of long-term growth

The authors have studied the effect of tool-making on man's evolution. They argue that growth is not exponential, but hyperbolic, and that a positive feedback exists linking technological development to population. Exploring the nature of this mechanism in order to improve the predictive capability of global models is the goal of the article.