Time series forecasting has played the key role in different industrial, including finance, traffic, energy, and healthcare domains. While existing literatures have designed many sophisticated architectures based on RNNs, GNNs, or Transformers, another kind of approaches multi-layer perceptrons (MLPs) are proposed with simple structure, low complexity, {superior performance}. However, most MLP-based methods suffer from point-wise mappings information bottleneck, which largely hinders...

Multivariate time series (MTS) forecasting has shown great importance in numerous industries. Current state-of-the-art graph neural network (GNN)-based methods usually require both networks (e.g., GCN) and temporal LSTM) to capture inter-series (spatial) dynamics intra-series (temporal) dependencies, respectively. However, the uncertain compatibility of two puts an extra burden on handcrafted model designs. Moreover, separate spatial modeling naturally violates unified spatiotemporal...

The ACEF score (age, creatinine, and left ventricular ejection fraction) the triglyceride-glucose (TyG) index have been identified as robust risk prediction models for adverse outcomes post-percutaneous coronary intervention (PCI) in atherosclerotic heart disease (CHD) patients. This study aimed to assess whether incorporating TyG enhances predictive ability of stratification CHD patients undergoing PCI. observational cohort enrolled 1248 diagnosed with who underwent PCI at Third People's...

The pattern set of a remote sensing image contains many kinds uncertainties. Uncertain information can create imperfect expressions for sets in various recognition algorithms, such as clustering algorithms. Methods based the fuzzy c-means algorithm manage some As soft methods, They are known to perform better on auto classification images than hard methods. However, if clusters different density and high order uncertainty, performance FCM may significantly vary depending choice fuzzifiers....

Noncontact human‐computer interaction has an important value in wireless sensor networks. This work is aimed at achieving accurate on a computer based auto eye control, using cheap webcam as the video source. A real‐time system state recognition, rough gaze estimation, and tracking proposed. Firstly, binary classification of states (opening or closed) carried SVM algorithm with HOG features input image. Second, appearance‐based estimation implemented simple CNN model. And head pose estimated...

In this paper a performance comparison of variety data preprocessing algorithms in remote sensing image classification is presented. These selected are principal component analysis (PCA) and three different independent analyses, ICA (Fast-ICA (Aapo Hyvarinen, 1999), Kernel-ICA (KCCA KGV (Bach & Jordan, 2002), EFFICA (Aiyou Chen Peter Bickel, 2003). were applied to imagery (1600×1197), obtained from Shunyi, Beijing. For classification, MLC method used for the raw preprocessed data. The...

The segmentation of buildings in remote-sensing (RS) images plays an important role monitoring landscape changes. Quantification these changes can be used to balance economic and environmental benefits most importantly, support the sustainable urban development. Deep learning has been upgrading techniques for RS image analysis. However, it requires a large-scale data set hyper-parameter optimization. To address this issue, concept “one view per city” is proposed explores use one parameter...

The preprocessing of remote sensing imagery (RSI) has great importance on the results classification. In this paper, algorithm fast independent component analysis (ICA) and its application to classification are presented, different parameter effect information extraction with ICA. for experiment is from areas, in time by several sensors. succession, a maximum likelihood estimation (MLE) supervised method used classify original images feature after As result, distinct characters images,...

Graph clustering algorithms with autoencoder structures have recently gained popularity due to their efficient performance and low training cost. However, for existing graph based on GCN or GAT, not only do they lack good generalization ability, but also the number of clusters clustered by such models is difficult determine automatically. To solve this problem, we propose a new framework called Clustering Masked Autoencoders (GCMA). It employs our designed fusion masking method coding graph....

We proposed a fast and fully automatic mosaic method for high-resolution satellite images. First, the overlapped rectangle is computed according to geographical locations of reference images feature points on both are extracted by scale-invariant transform (SIFT) algorithm only from region. Then, RANSAC used match Finally, two fused into seamlessly panoramic image simple linear weighted fusion or other method. The implemented in C++ language based OpenCV GDAL, tested Worldview-2...

In this paper, FastICA and KERNELICA algorithms their application as preprocessing for remote sensing imagery (RSI) classification are discussed, well a comparison between the two algorithms. Both of applied to TM RSI, obtained from Shunyi, Beijing, 1999. Then Maximum Likelihood Classification (MLC) method uniting ISODATA is used perform raw preprocessed data. The results show that use data gives more confident than those And ICA algorithms, on one side, both quite steady can get rid...

The result of linear target detection is advantageous to the content analysis and interpretation ASAR imagery can be regarded as matching reference well GIS platform input realize map automatic update. On foundation previous studies, method bridge based on Combined wavelet transformation mathematical morphology proposed according good local characteristic, multi-resolution effectiveness in this article, revised by combining fine characteristics different methods, which applied data Zhaoqing...

This paper focus on two phenomena that "same spectrum with different objects" and object spectra" in multispectral remote sensing image, propose a stepwise refinement classification method based multi-sensitive strategies. It's top-down, gradually hierarchical way of which combines advantages both supervised unsupervised classification: by analyzing the characteristic curve, cluster find out band combinations big differences as guidance classification; according to spectral characteristics...

Due to the scarcity of remote sensing image annotation data and low model generalization ability, automatically extracting buildings from different images remains a challenging problem. A domain adaption framework is proposed for building extraction tested on Inria aerial labeling dataset. Results show that beneficial improve accuracy in target domain.

Abstract Due to the influence of weather such as image blurring, license plate deformation, rain and snow, etc., in response problem high- precision automatic detection recognition non-HD plates complex environments, on basis more popular deep network, introduction dual attention mechanism for is proposed Model At_LPRNET, this basis, an integrated end-to-end network constructed that organically combined with batch The test results datasets chinese city parking dataset(CCPD) show At_LPRNET...