Shengliang Pu has a total of 32 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    HYPERSPECTRAL IMAGE CLASSIFICATION WITH LOCALIZED SPECTRAL FILTERING-BASED GRAPH ATTENTION NETWORK

    by S. Pu, Y. Song, Y. Chen, Y. Li, J. Zhang, Q. Lin, X. Zhu, H. Zeng, K. Liao, H. Yu, et al.

    Graph-based deep learning has been proved a promising approach that has an apparent superiority for learning graph data and modeling spatial topological relations between features. In particular, graph attention networks (GATs) are good at efficiently processing the graph-structured hyperspectral data by leveraging masked self-attention layers to address the known shortcomings of previous frameworks based on graph convolutions or their approximations. In this study, we proposed a novel approach that combines localized spectral filtering and GAT for the hyperspectral image classification task. First, we conducted unsupervised t-SNE (t-distributed stochastic neighbor embedding) manifold learning-based feature dimensionality reduction to create localized hyperspectral data cubes. Then, these feature cubes combined with localized adjacent matrices were fed into a shallow graph attention network in a supervised learning manner. Finally, we obtained credible classification results and promising classification performance in distinguishing diversified land covers through reducing the possible redundancy of spectral information and enhancing the expression of local spatial-spectral information. Experiments on two real hyperspectral data sets (that is, Indian Pines-A (IA) and Huanghekou (HH) data sets) demonstrated that the presented approach offers promising classification performance, that is, the GAT using t-SNE acquires superior performance than that of using PCA (principal component analysis), and also proves the great importance of combining spatial- and spectral information for hyperspectral image classification.


  • Open Access

    Review of: "Classification of regular and chaotic motions in Hamiltonian systems with deep learning"

    by Shengliang Pu

    In this manuscript, the authors take an interest in classifying types of motion on a time series data set coming from simple Hamiltonian systems and predicting the character of the dynamics on a model problem of the forced pendulum and an evolution p...

  • Open Access

    Spectral Graph Neural Networks with Manifold-Learning-Based Feature Extraction for Hyperspectral Image Classification

    by Shengliang Pu

    In this short article, we briefly retrospect the recent progress of spectral graph neural networks with manifold-learning-based feature extraction for hyperspectral image classification.

  • Open Access

    P-Learning

    by Shengliang Pu

    P-learning is a collection of machine learning techniques for training class descriptors on positive samples only.

  • Hyperspectral image classification with residual learning networks

    by Shengliang Pu, Lianru Gao, Yining Song, Yingyao Chen, Yating Li, Lingxin Luo, Guangyu Xu, Xiaowei Xie, Yunju Nie

    Hyperspectral imaging is particularly useful for per-pixel thematic classification by unique spectral signatures of landscape materials. Deep learning techniques such as convolutional neural networks have boosted the performance of image classification. Recently, several composite learning-based convolutional networks, i.e., deep residual networks (ResNets) and dense convolutional networks (DenseNets), have been presented to learn deep feature representation for image classification, and achieve high classification accuracies. In this paper, we present a fairly comparable architecture, including two kinds of modified residual learning networks with a shallow depth using small training data. First, we perform the extraction of key components from deep residual networks and dense convolutional networks, which is a set of composite learning structures with skip connections. Second, the plain convolutional neural networks (PNets) have been constituted by a stack of plain blocks that also have been placed in the presented network architecture as the baseline networks. Third, we make them as comparable as possible with the plain convolutional network structures, so that the more profound exploration and improvement could be further done. Finally, we wrap them together and design a comparable architecture. Experiments demonstrate that the presented residual learning networks show special characteristics for hyperspectral image classification, which have not been revealed before.

  • Remote Sensing Image Classification for Spatial Information Extraction of Panax notoginseng Fields

    by Shengliang Pu, Yining Song, Yingyao Chen, Yating Li, Lingxin Luo, Xiaowei Xie, Yunju Nie

    Chinese herbal medicine has played an important role in the treatment of the novel coronavirus patients. Machine learning-based remote-sensing techniques play a significant role in the quantitative resource inventory of materia medica resources, particularly to explore the monitoring abilities for sustainable utilization and biodiversity protection of the cultivated medicinal plant Panax notoginseng in the macrocosm. Until now, to the best knowledge, concrete planting patterns of Panax notoginseng are still poorly known. In this study, two popular supervised classifiers, i.e., support vector machine and artificial neural network, are employed to conduct mapping Panax notoginseng fields in Wenshan city, Yunnan province, China. It only targets a single class of interest, and there are 8,072 and 8,749 polygons extracted, whilst the planting areas of Panax notoginseng are estimated as 35.45 and 32.47 square kilometers achieved by SVM and ANN, respectively. Meanwhile, special concerns raised in terms of remote sensing mapping for modeling planting patterns and monitoring plant rotations of perennial Panax notoginseng cultivated under shade-net structures based on the experiments.

  • Open Access

    Lightweight Oriented Object Detection Using Multiscale Context and Enhanced Channel Attention in Remote Sensing Images

    by Qiong Ran, Qing Wang, Boya Zhao, Yuanfeng Wu, Shengliang Pu, Zijin Li

    Object detection is a focal point in remote sensing applications. Remote sensing images typically contain a large number of small objects and a wide range of orientations across objects. This results in great challenges to small object detection approaches based on remote sensing images. Methods directly employ channel relations with equal weights to construct information features leads to inadequate feature representation in complex image small object detection tasks. Multiscale detection methods improve the speed and accuracy of detection, while small objects themselves contain limited information, and the features are easily lost following down-sampling. During the detection, the feature images are independent across scales, resulting in a discontinuity at the detection scale. In this article, we propose the multiscale context and enhanced channel attention (MSCCA) model. MSCCA employs PeleeNet as the backbone network. In particular, the feature image channel attention is enhanced and the multiscale context information is fused with multiscale detection methods to improve the characterization ability of the convolutional neural network. The proposed MSCCA method is evaluated on two real datasets. Results show that for 512 × 512 input images, MSCCA was able to achieve 80.4% and 94.4% mAP on the DOTA and NWPU VHR-10, respectively. Meanwhile, the model size of MSCCA is 21% smaller than that of its predecessor. MSCCA can be considered as a practical lightweight oriented object detection model in remote sensing images.


  • Spatial Information Extraction of Panax Notoginseng Fields Using Multi-algorithm and Multi-sample Strategy-Based Remote Sensing Techniques

    by Shengliang Pu, Yining Song, Yingyao Chen, Yating Li, Lingxin Luo, Guangyu Xu, Xiaowei Xie, Yunju Nie

  • Open Access

    Hyperspectral Image Classification with Localized Graph Convolutional Filtering

    by Shengliang Pu, Yuanfeng Wu, Xu Sun, Xiaotong Sun

    The nascent graph representation learning has shown superiority for resolving graph data. Compared to conventional convolutional neural networks, graph-based deep learning has the advantages of illustrating class boundaries and modeling feature relationships. Faced with hyperspectral image (HSI) classification, the priority problem might be how to convert hyperspectral data into irregular domains from regular grids. In this regard, we present a novel method that performs the localized graph convolutional filtering on HSIs based on spectral graph theory. First, we conducted principal component analysis (PCA) preprocessing to create localized hyperspectral data cubes with unsupervised feature reduction. These feature cubes combined with localized adjacent matrices were fed into the popular graph convolution network in a standard supervised learning paradigm. Finally, we succeeded in analyzing diversified land covers by considering local graph structure with graph convolutional filtering. Experiments on real hyperspectral datasets demonstrated that the presented method offers promising classification performance compared with other popular competitors.


  • Open Access

    Hyperspectral Image Classification with Capsule Network Using Limited Training Samples

    by Fei Deng, Shengliang Pu, Xuehong Chen, Yusheng Shi, Ting Yuan, Shengyan Pu

    Deep learning techniques have boosted the performance of hyperspectral image (HSI) classification. In particular, convolutional neural networks (CNNs) have shown superior performance to that of the conventional machine learning algorithms. Recently, a novel type of neural networks called capsule networks (CapsNets) was presented to improve the most advanced CNNs. In this paper, we present a modified two-layer CapsNet with limited training samples for HSI classification, which is inspired by the comparability and simplicity of the shallower deep learning models. The presented CapsNet is trained using two real HSI datasets, i.e., the PaviaU (PU) and SalinasA datasets, representing complex and simple datasets, respectively, and which are used to investigate the robustness or representation of every model or classifier. In addition, a comparable paradigm of network architecture design has been proposed for the comparison of CNN and CapsNet. Experiments demonstrate that CapsNet shows better accuracy and convergence behavior for the complex data than the state-of-the-art CNN. For CapsNet using the PU dataset, the Kappa coefficient, overall accuracy, and average accuracy are 0.9456, 95.90%, and 96.27%, respectively, compared to the corresponding values yielded by CNN of 0.9345, 95.11%, and 95.63%. Moreover, we observed that CapsNet has much higher confidence for the predicted probabilities. Subsequently, this finding was analyzed and discussed with probability maps and uncertainty analysis. In terms of the existing literature, CapsNet provides promising results and explicit merits in comparison with CNN and two baseline classifiers, i.e., random forests (RFs) and support vector machines (SVMs).