Spatial Weight Research Articles

Robust Localization-Guided Dual-Branch Network for Camouflaged Object Segmentation

The existence of camouflage targets is widespread in the natural world, as they blend seamlessly or closely resemble their surrounding environment, making it difficult for the human eye to identify them accurately. In camouflage target segmentation, challenges often arise from the high similarity between the foreground and background, resulting in segmentation errors, imprecise edge detection, and overlooking of small targets. To address these issues, this paper presents a robust localization-guided dual-branch network for the recognition of camouflaged targets. Two crucial branches, i.e., a localization branch and an overall refinement branch are designed and incorporated. The localization branch achieves accurate preliminary localization of camouflaged targets by incorporating the robust localization module, which integrates different high-level feature maps in a partially decoded manner. The overall refinement branch optimizes segmentation accuracy based on the output predictions of the localization branch. Within this branch, the edge refinement module is devised to effectively reduce false negative and false positive interference. By conducting context exploration on each feature layer from top to bottom, this module further enhances the precision of target edge segmentation. Additionally, our network employs five jointly trained output prediction maps and introduces attention-guided heads for diverse prediction maps in the overall refinement branch. This design adjusts the spatial positions and channel weights of different prediction maps, generating output prediction maps based on the emphasis of each output, thereby further strengthening the perception and feature representation capabilities of the model. To improve its ability to generate highly confident and accurate prediction candidate regions, tailored loss functions are designed to cater to the objectives of different prediction maps. We conducted experiments on three publicly available datasets for camouflaged object detection to assess our methodology and compared it with state-of-the-art network models. On the largest dataset COD10K, our method achieved a Structure-measure of 0.827 and demonstrated superior performance in other evaluation metrics, outperforming recent network models.

Multi-View Scene Classification Based on Feature Integration and Evidence Decision Fusion

Leveraging multi-view remote sensing images in scene classification tasks significantly enhances the precision of such classifications. This approach, however, poses challenges due to the simultaneous use of multi-view images, which often leads to a misalignment between the visual content and semantic labels, thus complicating the classification process. In addition, as the number of image viewpoints increases, the quality problem for remote sensing images further limits the effectiveness of multi-view image classification. Traditional scene classification methods predominantly employ SoftMax deep learning techniques, which lack the capability to assess the quality of remote sensing images or to provide explicit explanations for the network’s predictive outcomes. To address these issues, this paper introduces a novel end-to-end multi-view decision fusion network specifically designed for remote sensing scene classification. The network integrates information from multi-view remote sensing images under the guidance of image credibility and uncertainty, and when the multi-view image fusion process encounters conflicts, it greatly alleviates the conflicts and provides more reasonable and credible predictions for the multi-view scene classification results. Initially, multi-scale features are extracted from the multi-view images using convolutional neural networks (CNNs). Following this, an asymptotic adaptive feature fusion module (AAFFM) is constructed to gradually integrate these multi-scale features. An adaptive spatial fusion method is then applied to assign different spatial weights to the multi-scale feature maps, thereby significantly enhancing the model’s feature discrimination capability. Finally, an evidence decision fusion module (EDFM), utilizing evidence theory and the Dirichlet distribution, is developed. This module quantitatively assesses the uncertainty in the multi-perspective image classification process. Through the fusing of multi-perspective remote sensing image information in this module, a rational explanation for the prediction results is provided. The efficacy of the proposed method was validated through experiments conducted on the AiRound and CV-BrCT datasets. The results show that our method not only improves single-view scene classification results but also advances multi-view remote sensing scene classification results by accurately characterizing the scene and mitigating the conflicting nature of the fusion process.

Direct and spillover effects of short‐ and long‐term land pricing drivers: Evidence from Italian districts, 1992−2019

AbstractThe global economic and food crisis has increased the demand for land and rekindled the interest in farmland market investments worldwide. This study explores the Italian farmland market, investigating its main influencing factors from 1992 to 2019 using a spatial econometric framework. Traditional land characteristics and location‐specific agricultural factors, as well as non‐agricultural factors, are assessed. The average level of farmland prices and their growth are analyzed by modeling the potential types of spatial interactions, and the results are corroborated by considering different configurations of spatial weight matrices. The results show that farmland markets are influenced by land's current net returns as well as by its potential alternative uses. Therefore, factors considered external to the agricultural dimension, such as population pressure, climate change, and speculative expectations, increasingly shape farmland prices.

Combining Swin Transformer and Attention-Weighted Fusion for Scene Text Detection

The existing text detection algorithms based on Convolutional Neural Networks (CNN) commonly have the problems of insufficient receptive fields and inadequate extraction of spatial positional information, which limit their ability to detect large-scale variation text instances, long-distance and wide-spaced text instances as well as effectively distinguish complex background textures. To address the above problems, in this paper, a scene text detection algorithm combining Swin Transformer and attention-weighted fusion is proposed. Firstly, an attention-weighted fusion (AWF) module is proposed, which embeds a modified coordinate attention module (CAM) in the feature pyramid network (FPN). This module learns spatial positional weights of foreground information in different-scale features while suppressing redundant background information. As a result, the fused features are more focused on the text regions, enhancing the localization ability for text regions and boundaries. Secondly, the window-based self-attention mechanism of the Swin Transformer is utilized to achieve global feature perception on the fused features of the pyramid network. This compensates for the insufficient receptive fields of CNN and enhances the representation capability of global contextual features, thereby further improving the performance of text detection. Experimental results demonstrate that the proposed algorithm achieves competitive performance on three public datasets, namely ICDAR2015, MSRA-TD500, and Total-Text, with F-measure reaching 87.9%, 91.4%, and 86.7%, respectively. Code is available at: https://github.com/xgli411/ST-AWFNet.

Study on the spillover effect of digital economy development on CO2 emissions

To study the spillover effect of the digital economy development on carbon emissions, this study innovatively constructs different spatial weight matrices, based on 2011–2020 panel data covering 30 Chinese provinces, and it explores the direct spillovers, conducted spillovers, and spillovers from different spillover channels, such as human capital, service industry development, and information development of digital economy development on carbon emissions through the spatial Durbin model combined with a mediating effect model. The results show that there is significant spatial heterogeneity in digital economy development; in terms of regions, the eastern region has the highest average development level and the central region has the highest average annual growth rate. Digital economy development can directly suppress carbon emissions, and it can also indirectly suppress carbon emissions by driving technological innovation and optimizing the energy consumption structure, and there exists a spatial spillover effect. Under human capital, service industry development and information development matrices, the spatial spillover effect of digital economy development on carbon emissions is significantly negative. Regions with the same level of information development are more likely to exert a spatial spillover effect of digital economy development on carbon emissions.

Spatial–temporal feature-based End-to-end Fourier network for 3D sign language recognition

Most dynamic sign word misclassifications are caused by redundant spatial–temporal (SPT) feature pruning that lacks language semantic and temporal dependencies. SPT feature recognition is one of the important aspects for the evaluation of the misclassification of dynamic sign words. The redundant pruning of SPT feature space influences the language model of sign confusion, model complexity, and SPT feature similarity. The purpose of this article is to develop a new multi-scale SPT feature-based dynamic sign word recognition approach via a low-cost feature selection method (FS) and End-to-end Fourier convolution neural network (EFCNN). Instead of a sensor fusion technique for obtaining frame position alignment, in the EFCNN, new 3D frame position and coordinates are determined using a pixel weighting and alignment function of the first and succeeding 25 spatial intensities of the 3D video changes across hand motion. The new spatial weight and the original spatial coordinates are fused and truncated in the Fourier domain. We generate the temporal dependence of the fused features. A feature selection known as the FS-EFCNN is introduced to select compact features with a preserved language meaning. Five state-of-the-art feature selection methods, namely Infinite FS (InFS), Relief FS, Fisher, MIM, ILFS, and ensemble FS-EFCNN were deployed to guide and optimize the learning performance of EFCNN. The experimental result analysis highlighted the improved results of the FS-EFCNN method with the best accuracy of 99.86%, 99.89%, and 90.69% on 3D American Sign Language, British Sign Language, and Greek Sign Language data sets, respectively.

Study on the impact of digital finance on agricultural carbon emissions from a spatial perspective: an analysis based on provincial panel data

Abstract China is committed to achieving the goal of ‘carbon peaking and carbon neutrality’, which requires reducing carbon emissions from various sectors, especially agriculture. Agriculture is a major source of greenhouse gas emissions, and its carbon intensity is influenced by many factors, including the development of digital finance. Digital finance, as an important driver of economic growth and technological innovation, can affect the carbon emissions of agricultural production and consumption through various channels. However, the existing literature has not paid enough attention to the spatial effects of digital finance on agricultural carbon emissions, which may lead to biased estimates and policy implications. Therefore, this paper aims to fill this gap by using the Spatial Durbin Model to analyze the impact of digital finance on agricultural carbon emissions from a spatial perspective, based on the provincial panel data of 28 provinces (municipalities and autonomous regions) in China from 2011 to 2020. The main findings are as follows: (1) The development of digital finance has a positive and significant effect on China’s agricultural carbon emissions, and this effect is robust to different spatial weight matrices. (2) The impact mechanism of digital finance on agricultural carbon emissions is mainly through affecting the spatial spillover of carbon emissions and the cross-regional financial coordination. Based on these results, this paper provides some policy suggestions for achieving the goal of ‘carbon peaking and carbon neutrality’ in China’s agricultural sector.

Digitalization, Spillover and Environmental, Social, and Governance Performance: Evidence From China

There is growing interest in the relationship between digitalization and environmental, social, and governance (ESG) performance, but existing research focuses on the one-way relationship and ignores the two-way mechanism. Based on a sample of 3335 listed companies in China in 2020, this study adopts a spatial simultaneous equation model to investigate the bidirectional mechanism between companies’ digitalization and ESG performance. The results show that digitalization and ESG performance have a significant positive two-way mechanism; digitalization enhances ESG performance, while ESG performance promotes digitalization. The results also indicate a significant positive intra-industry spillover effect for both digitization and ESG performance. Further research shows that the relationship between digitization and ESG performance does not differ depending on the definition of spatial weights, however, the spillover effects do differ depending on the definition of spatial weights.

Universal REFoCUS beamforming with spatial weighting

Universal REFoCUS beamforming with spatial weighting

Salient Object Detection With Edge-Guided Learning and Specific Aggregation

Recently, the performance of salient object detection (SOD) has been significantly improved by utilizing edge information for auxiliary training. However, the extraction and utilization of edge cues and multi-level feature fusion are still two issues in existing edge-aware models. In this paper, we devise a novel SOD network with edge-guided learning and specific aggregation, named ELSA-Net, to cooperatively address these two issues. First, we propose the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">edge-guided learning strategy</i> , which utilizes edge cues as low-level guidance to improve saliency prediction. Specifically, we design a two-stream model that uses a saliency branch and an edge branch to detect the interior and the boundary of salient objects, respectively. Then, an edge-guided interaction module (EGI) is further designed to achieve feature enhancement by embedding edge information into the saliency branch as the spatial weights. In addition, two <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">specific aggregation modules</i> are proposed for the progressive fusion of multi-level features in the above two streams, thus making full use of semantic and detailed information. The high-level interactive fusion module (HIF) leverages the correlation between two deeper features to obtain more powerful global contexts. And the low-level weighted fusion module (LWF) focuses on the complement of fine information by selectively integrating input features. Extensive experiments show that the proposed approach outperforms 19 state-of-the-art methods on five datasets, which validates its effectiveness both quantitatively and qualitatively.

Modelling the Count of Diphtheria in West Java in 2022 using Geographically Weighted Zero-Inflated Poisson Regression

The number of death cases caused by infectious diseases, that is to say diphtheria, especially in West Java so that it is categorized as an Extraordinary Event (KLB) generally attacks from the age of toddlers to the elderly. To model diphtheria cases in West Java in 2022 containing over dispersion using Geographically Weighted Zero-Inflated Poisson Regression (GWZIPR). Parameter estimation uses Expectation Maximization (EM) algorithm with spatial weight using Adaptive Gaussian Kernel. The resulting models include a Poisson state model that shows the chance of diphtheria cases occurring with relevant factors of the percentage within the population who apply Hygienic and Healthful Lifestyle (PHBS), who have been vaccinated against DPT-HB3, the number of health workers at Puskesmas, and the percentage of active posyandu and a Zero state model that shows the chance of no diphtheria cases occurring with relevant factors of the percentage who have been vaccinated against DPT-HB3, the number of health workers at Puskesmas, and the percentage of active posyandu.

Does the Collaborative Agglomeration of Logistics and Manufacturing have an Inhibitory Effect on Carbon Emissions? Based on the Spatial Econometric Analysis of Panel Data of 30 Provinces in China

In this paper, panel data from 30 provinces in the Chinese Mainland (excluding Tibet, Hong Kong, Macao, and Taiwan) spanning from 2006 to 2022 is utilized to investigate the relationship between logistics industry agglomeration, manufacturing industry agglomeration, and carbon emissions using spatial econometric analysis methodologies. Specifically, the study focuses on investigating the relationship between the synergistic agglomeration of the logistics and manufacturing industry, and their impact on carbon emissions using the Spatial Durbin Model. The research findings indicate that an increase in manufacturing industry agglomeration significantly reduces carbon emissions, an increase in logistics industry agglomeration result in higher carbon emissions, and the synergistic agglomeration of the two industries plays a significant role in mitigating regional carbon emissions through spatial and geographical mechanisms. Robustness tests, which include the control of additional variables and the use of alternative spatial weight matrices, further support the above conclusions.

Spartus: A 9.4 TOp/s FPGA-Based LSTM Accelerator Exploiting Spatio-Temporal Sparsity.

Long short-term memory (LSTM) recurrent networks are frequently used for tasks involving time-sequential data, such as speech recognition. Unlike previous LSTM accelerators that either exploit spatial weight sparsity or temporal activation sparsity, this article proposes a new accelerator called "Spartus" that exploits spatio-temporal sparsity to achieve ultralow latency inference. Spatial sparsity is induced using a new column-balanced targeted dropout (CBTD) structured pruning method, producing structured sparse weight matrices for a balanced workload. The pruned networks running on Spartus hardware achieve weight sparsity levels of up to 96% and 94% with negligible accuracy loss on the TIMIT and the Librispeech datasets. To induce temporal sparsity in LSTM, we extend the previous DeltaGRU method to the DeltaLSTM method. Combining spatio-temporal sparsity with CBTD and DeltaLSTM saves on weight memory access and associated arithmetic operations. The Spartus architecture is scalable and supports real-time online speech recognition when implemented on small and large FPGAs. Spartus per-sample latency for a single DeltaLSTM layer of 1024 neurons averages 1 μ s. Exploiting spatio-temporal sparsity on our test LSTM network using the TIMIT dataset leads to 46 × speedup of Spartus over its theoretical hardware performance to achieve 9.4-TOp/s effective batch-1 throughput and 1.1-TOp/s/W power efficiency.

The impact of digital finance on the urban-rural income gap: an empirical analysis based on spatial econometric modeling

Abstract China’s economy has stepped into the new normal development stage. Coordinating urban and rural economic development and continuously narrowing the income gap has become a more urgent strategic task, and digital finance is an important engine for economic development and transformation and upgrading, is bound to have a far-reaching impact on the narrowing of the urban-rural income gap. This paper first explores the essential relationship between digital finance and regional economic development and lists the important roles played by digital finance in regional economic development. The entropy power Topsis method is used to select variable indicators and construct the index system. Combining the Moran index, spatial spillover effect, and spatial panel regression model, the regional economic distance is used as the spatial weight matrix, and the corresponding elemental expressions are derived to construct the spatial econometric model of the urban-rural income gap by digital finance, and the model parameters are estimated by the great likelihood method. The model is applied to 31 provinces in China to measure their relevant variables from 2012 to 2022 and analyze the impact of digital finance on the urban-rural income gap from the perspectives of spatial autocorrelation and threshold regression. Through the experiment, it can be seen that the digital finance indexes of the eastern, central, and western regions are all negative, and they are significant at the 5%, 10%, and 10% levels, respectively, which indicates that the digital finance in these three regions stimulates the development of the financial industry while narrowing the urban-rural income gap. When the level of economic development crosses the first threshold value of 10.4365, the regression coefficient of digital finance decreases from −0.0348 to −0.0487, which again confirms that digital finance can reduce the urban-rural income gap.

Comparison of distance-based spatial weight matrix in modeling Internet signal strengths in Tasikmalaya regency using logistic spatial autoregressive model

To ensure that national development objectives in rural areas are achieved evenly and sustainably, the Government of Indonesia applies the principles of Village Sustainable Development Goals (SDGs), which are derivative programs of SDGs. One of the indicators in measuring the progress and independence of villages in Indonesia is the availability of cellular phone signal access. Cellular phone signals have a vital role because most internet users in Indonesia rely on mobile data connections from cellular operators. However, the signal emitted by a provider tower has a limited range. According to the data of the Developing Villages Index in 2022, Tasikmalaya Regency is one of the regencies with the highest number of villages that have weak signal strength in West Java Province, Indonesia. To examine the effect of distance and height difference between the placement of the nearest provider tower and the location of the Village Office on the internet signal strength category in Tasikmalaya Regency, Logistic Spatial Autoregressive modeling is needed. In this study, the Bayesian Markov-Chain Monte Carlo estimation method was used, because it has advantages in flexibility and computational efficiency. In spatial modeling, there is a spatial weight matrix determined by the researcher’s understanding of the observed phenomenon. The variable observed in this study is signal strength, which has an orientation at a distance. However, there are several types of distance-based spatial weight matrices, such as K-nearest neighbor, radial distance, power distance, and exponential distance. To determine the most suitable distance-based spatial weight matrix in internet signal strength modeling, the four (4) weight matrices were compared based on the goodness of fit measure models, calculated from the confusion matrix. The results of the analysis showed that the radial distance weight matrix with a threshold distance of d = 1.7km is the most suitable use of distance-based spatial weight matrix in internet signal modeling in Tasikmalaya Regency. The weight matrix exerted a positive spatial autocorrelation effect of 57.141%. In addition, the height difference factor between the location of the provider tower with the location of the village office has a greater effect than the horizontal distance.

Study on the historical landscape evolution of the core reserve of Guangfu Ancient City based on computer simulation

Abstract In the conservation of the historical landscape within the Guangfu Ancient City Core Protection Area, traditional descriptions via drawings and textual narratives often exhibit subjective biases. To address this, a theoretical model for the historical architectural landscape has been developed to objectively measure and characterize the spatial patterns and morphologies of the landscape. This study introduces Exploratory Spatial Data Analysis (ESDA) techniques, utilizing spatial weights and spatial correlation measures as primary analytical tools. Additionally, meta-cellular automata are employed to simulate the spatial and temporal evolution of the historical landscape. Geographical Information Systems (GIS) and Remote Sensing (RS) technologies have been leveraged to establish a spatial database and facilitate the visualization of the historical landscape’s evolutionary processes. This integrated model and technological approach enable a detailed analysis of the architectural landscape characteristics and the spatial structure’s evolution within the study area. The findings reveal distinct spatial distribution characteristics, with an undulation index ranging from 5 to 20 and a predominant index over 7, accounting for 15.67% of the land. Based on these insights, the study proposes strategic interventions to safeguard the historical integrity of Guangfu Ancient City’s core conservation area.

M 2 FTrans: Modality-Masked Fusion Transformer for Incomplete Multi-Modality Brain Tumor Segmentation.

Brain tumor segmentation is a fundamental task and existing approaches usually rely on multi-modality magnetic resonance imaging (MRI) images for accurate segmentation. However, the common problem of missing/incomplete modalities in clinical practice would severely degrade their segmentation performance, and existing fusion strategies for incomplete multi-modality brain tumor segmentation are far from ideal. In this work, we propose a novel framework named M 2 FTrans to explore and fuse cross-modality features through modality-masked fusion transformers under various incomplete multi-modality settings. Considering vanilla self-attention is sensitive to missing tokens/inputs, both learnable fusion tokens and masked self-attention are introduced to stably build long-range dependency across modalities while being more flexible to learn from incomplete modalities. In addition, to avoid being biased toward certain dominant modalities, modality-specific features are further re-weighted through spatial weight attention and channel- wise fusion transformers for feature redundancy reduction and modality re-balancing. In this way, the fusion strategy in M 2 FTrans is more robust to missing modalities. Experimental results on the widely-used BraTS2018, BraTS2020, and BraTS2021 datasets demonstrate the effectiveness of M 2 FTrans, outperforming the state-of-the-art approaches with large margins under various incomplete modalities for brain tumor segmentation. Code is available at https://github.com/Jun-Jie-Shi/M2FTrans.

ВИКОРИСТАННЯ ПРОСТОРОВИХ ВАГОВИХ МАСОК З НЕЗБАЛАНСОВАНИМИ ДАТАСЕТАМИ ДЛЯ ЗАДАЧІ СЕГМЕНТАЦІЇ СУПУТНИКОВИХ ЗНІМКІВ

This study addresses the issue of imbalanced datasets in satellite image segmentation tasks, where machine learning models often neglect minority classes in favor of majority ones. We propose using spatial weight masks for the loss function computation to take into account the reliability score of individual pixels. This approach enhances segmentation quality, significantly improving metrics for minority classes. Additionally, a dataset augmentation method using generative adversarial networks (GANs) is explored, showing slight improvements in recognizing less represented crop types in the dataset, and it is compared with the proposed method. The simultaneous usage of weighted masks and generative networks is investigated. Keywords: dataset quality assessment, imbalanced datasets, classification, segmentation, generative adversarial networks, training data generation.

Performance degradation of steel rebar considering corrosion spatial variability in concrete

The corrosion of steel bars is the main issue affecting the durability of reinforced concrete structures, which is nonuniform owing to the material properties and external environmental effects. In this study, the performance degradation of steel rebars in concrete is estimated considering corrosion spatial variability. With the aid of a three-dimensional measurement technique, a finite element model based on the scanning result was established to examine the corrosion effect on the tensile behaviour of the corroded bars, which was validated with the experimental results and digital image correlation observation. To assess the deformation behaviour of the corroded steels, a semi-analytical model was developed based on the cross-sectional area distribution and material properties. Subsequently, the spatial weight loss distributions with different degrees of corrosion were generated using a random field model. The model parameters were determined by combining engineering judgement and estimation of the experimental results. Finally, probabilistic model of degraded performance was established using the deterministic assessment and simulated steel weight loss distribution.

Lightweight 2D Human Pose Estimation Based on Joint Channel Coordinate Attention Mechanism

Traditional human pose estimation methods typically rely on complex models and algorithms. Lite-HRNet can achieve an excellent performance while reducing model complexity. However, its feature extraction scale is relatively single, which can lead to lower keypoints’ localization accuracy in crowded and complex scenes. To address this issue, we propose a lightweight human pose estimation model based on a joint channel coordinate attention mechanism. This model provides a powerful information interaction channel, enabling features of different resolutions to interact more effectively. This interaction can solve the problem of human pose estimation in complex scenes and improve the robustness and accuracy of the pose estimation model. The introduction of the joint channel coordinate attention mechanism enables the model to more effectively retain key information, thereby enhancing keypoints’ localization accuracy. We also redesign the lightweight basic module using the shuffle module and the joint channel coordinate attention mechanism to replace the spatial weight calculation module in the original Lite-HRNet model. By introducing this new module, we not only improve the network calculation speed and reduce the number of parameters of the entire model, but also ensure the accuracy of the model, thereby achieving a balance between performance and efficiency. We compare this model with current mainstream methods on the COCO and MPII dataset. The experimental results show that this model can effectively reduce the number of parameters and computational complexity while ensuring high model accuracy.