Hyperspectral Image Classification With Spectral and Spatial Graph Using Inductive Representation Learning Network

Pan, Yang; Tong, Lei; Qian, Bin; Gao, Zheng; Yu, Jing; Xiao, Chuangbai

doi:10.1109/jstars.2020.3042959

Cited by 26 publications

(7 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The connection state of nodes can change, which effectively improves the generalization ability and scalability of the model in MANETs. In addition, compared with embedding methods such as GCN and GraphSAGE 33 , FastGCN can reduce the time complexity and improve the efficiency of the algorithm by using Monte Carlo method to approximate the computation of the convolution and loss function by node sampling. In FastGCN, the simplest way to sample nodes is to use a uniform distribution for sampling.…”

Section: Methodsmentioning

confidence: 99%

A link prediction method for MANETs based on fast spatio-temporal feature extraction and LSGANs

Shao

Wang

Liu

et al. 2022

Sci Rep

View full text Add to dashboard Cite

Link prediction aims to learn meaningful features from networks to predict the possibility of topology. Most of the existing research on temporal link prediction is mainly aimed at networks with slow topology changes. They ignore the information of topology interval and link duration. This paper proposes a link prediction model named FastSTLSG. It can automatically analyze the features of the topology in a unified framework to effectively capture the spatio-temporal correlation of Mobile Ad Hoc Networks. First, we regard the changing topology as a chaotic system, transform it into a series of static snapshots based on the autocorrelation function; Next, the fast graph convolutional network efficiently analyses the topological relationships between nodes and reduces the computational complexity by importance sampling. Then, the gate recurrent unit captures the temporal correlation between snapshots. Finally, the fully connected layer reconstructs the topological structure. In addition, we take full advantage of least squares generative adversarial networks to further improve the performance of generator to obtain high-quality link prediction results. Extensive experiments on different datasets show that our FastSTLSG model obtains higher prediction accuracy compared with existing baseline models.

show abstract

Section: Methodsmentioning

confidence: 99%

A link prediction method for MANETs based on fast spatio-temporal feature extraction and LSGANs

Shao

Wang

Liu

et al. 2022

Sci Rep

View full text Add to dashboard Cite

show abstract

“…Furthermore, techniques like Deep Recurrent Neural Networks [47], 3D deep learning frameworks [48], [49], [50], [51], Cascaded Recurrent Neural Networks [31], and Multi-Layer Perceptrons (MLP) [46] emerged as prominent contenders in HSI classification. Hybrid approaches such as Spiking Neural Networks (SNN) [52], 1D CNN [28], Morphological Convolutional Neural Networks (MCNN) [30], S2GraphSage [29], RLSBSA [26], and 3DHyperGamo [27] models have broken previous accuracy records in predicting HSI classes.…”

Section: Convolution Neural Networkmentioning

confidence: 99%

Deep Feature Segmentation Model Driven by Hybrid Convolution Network for Hyper Spectral Image Classification

K Ghotekar,

Shaw,

Rout

2024

IJCDS

View full text Add to dashboard Cite

Hyper-spectral image (HSI) classification can support different applications, such as agriculture, military, city planning, land utilization, and identifying distinct regions. It is treated as a crucial topic in the research community. Recent advancements in convolution neural networks (CNN) have shown the unique capability of extracting meaningful features and classification. However, CNN works with square images with fixed dimensions and cannot extract local information of images having distinct geometric variations with context and content relationships; hence, there is a scope for improvement in correctly identifying class boundaries. Encouraged by the facts, we propose an HSI feature segmentation model by the hybrid convolution network (GCNN-RESNET152) for the HSI classification. First, pre-trained CNN on ImageNet is used to obtain the multi-layer feature. Second, the 3D discrete wavelet transform image is fed into the graph convolution network GCN model to gain patch-to-patch correlations feature maps. Finally, the features are integrated using the concatenation method of the three-weighted coefficients. Finally, the linear classifier is used to predict the semantic classes of pixel HSI. The proposed model is tested on four benchmark datasets: Houston University (HU), Indian Pines (IP), Kennedy Space Station (KSS), and Pavia University (PU). The result is compared with state-of-art algorithms and is superior in terms of overall, average, and kappa accuracy.

show abstract

“…During the last years, GCN s have found extensive use in a diverse range of applications, including citation and social networks [ 19 , 26 ], graph-to-sequence learning tasks in natural language processing [ 27 ], molecular/compound graphs [ 28 ], and action recognition [ 29 ]. Considerable research has been conducted on the classification of remotely sensed hyperspectral images [ 30 , 31 , 32 ], mainly due to the capabilities of GCN s to capture both the spatial contextual information of pixels, as well as the long-range relationships of distant pixels in the image. Another domain of application is the forecasting of traffic features in smart transportation networks [ 33 , 34 ].…”

Section: Introductionmentioning

confidence: 99%

Dense Multi-Scale Graph Convolutional Network for Knee Joint Cartilage Segmentation

Chadoulos,

Tsaopoulos,

Symeonidis

et al. 2024

Bioengineering

View full text Add to dashboard Cite

In this paper, we propose a dense multi-scale adaptive graph convolutional network (DMA-GCN) method for automatic segmentation of the knee joint cartilage from MR images. Under the multi-atlas setting, the suggested approach exhibits several novelties, as described in the following. First, our models integrate both local-level and global-level learning simultaneously. The local learning task aggregates spatial contextual information from aligned spatial neighborhoods of nodes, at multiple scales, while global learning explores pairwise affinities between nodes, located globally at different positions in the image. We propose two different structures of building models, whereby the local and global convolutional units are combined by following an alternating or a sequential manner. Secondly, based on the previous models, we develop the DMA-GCN network, by utilizing a densely connected architecture with residual skip connections. This is a deeper GCN structure, expanded over different block layers, thus being capable of providing more expressive node feature representations. Third, all units pertaining to the overall network are equipped with their individual adaptive graph learning mechanism, which allows the graph structures to be automatically learned during training. The proposed cartilage segmentation method is evaluated on the entire publicly available Osteoarthritis Initiative (OAI) cohort. To this end, we have devised a thorough experimental setup, with the goal of investigating the effect of several factors of our approach on the classification rates. Furthermore, we present exhaustive comparative results, considering traditional existing methods, six deep learning segmentation methods, and seven graph-based convolution methods, including the currently most representative models from this field. The obtained results demonstrate that the DMA-GCN outperforms all competing methods across all evaluation measures, providing DSC=95.71% and DSC=94.02% for the segmentation of femoral and tibial cartilage, respectively.

show abstract

Hyperspectral Image Classification With Spectral and Spatial Graph Using Inductive Representation Learning Network

Cited by 26 publications

References 38 publications

A link prediction method for MANETs based on fast spatio-temporal feature extraction and LSGANs

A link prediction method for MANETs based on fast spatio-temporal feature extraction and LSGANs

Deep Feature Segmentation Model Driven by Hybrid Convolution Network for Hyper Spectral Image Classification

Dense Multi-Scale Graph Convolutional Network for Knee Joint Cartilage Segmentation

Contact Info

Product

Resources

About