Representative band selection for hyperspectral image classification

Yang, Ronglu; Su, Lifan; Zhao, Xibin; Wan, Hai; Sun, Jun

doi:10.1016/j.jvcir.2017.02.002

Cited by 51 publications

(24 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Many hyperspectral data analysis techniques use dimensionality reduction as a pre-processing step, which aims to remove the redundant spectral information while preserving only critical information for subsequent processing. Hyperspectral dimensionality reduction can be achieved through feature extraction or band selection [44], [45]. Feature extraction methods like Principal Component Analysis (PCA) and Minimum Noise Fraction (MNF) transform the original data into reduced feature spaces by means of different criteria, whereas band selection aims to select a small subset of hyperspectral bands to reduce the burden of heavy computations [46].…”

Section: Problem Statement and Proposed Algorithmmentioning

confidence: 99%

Entropy-Based Convex Set Optimization for Spatial–Spectral Endmember Extraction From Hyperspectral Images

Shah

Zaveri

Trivedi

2020

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Spectral unmixing is an important problem for remotely sensed hyperspectral data exploitation. Automatic spectral unmixing can be viewed as a three-stage problem, where the first stage is subspace identification, the next one is endmember extraction, and the final one is abundance estimation. In this sequence, endmember extraction is the most challenging problem. Many researchers have attempted to extract endmembers from hyperspectral images using spectral information only. However, it is well known that the inclusion of spatial information can improve the endmember extraction task. In this paper, we introduce a new endmember extraction algorithm that exploits both spectral and spatial information. A main innovation of the proposed algorithm is that spatial information is exploited using entropy, while spectral information is exploited using convex set optimization. In the literature, none of the spatial-spectral algorithms has used entropy as spatial information. The inclusion of this entropy-based spatial information improves the accuracy of the endmember extraction process. The results obtained by the proposed algorithm are compared (using a variety of metrics) with those obtained by other state-of-the-art methods, using both synthetic and real datasets. Our experimental results demonstrate that the proposed algorithm outperforms many available algorithms.

show abstract

Section: Problem Statement and Proposed Algorithmmentioning

confidence: 99%

Entropy-Based Convex Set Optimization for Spatial–Spectral Endmember Extraction From Hyperspectral Images

Shah

Zaveri

Trivedi

2020

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

show abstract

“…Then, we randomly choose 10% of the data to train the SVM classifier in all the methods and obtain the classification accuracy of all methods in different selected bands. Second, after band selection was performed by all the methods, according to [49], [50] which used 20% of data to train SVM, we also choose another 20% of the data to train a new SVM classifier compared with STMIGR [49], improved sparse subspace clustering (ISSC) [50], and WaLuDi, and our methods included GradHM+AS, GradHM+TS, Guided-GradHM+AS, and Guided-GradHM+TS. Third, the CNN model for the Indian Pines data should be trained by different proportions samples to obtain GradHM and Guided-GradHM.…”

Section: B Experimental Setupmentioning

confidence: 99%

Band Selection via Explanations From Convolutional Neural Networks

et al. 2020

View full text Add to dashboard Cite

Band Selection is a research hotspot in the field of hyperspectral imaging (HSI) processing. This paper proposes a method that selects bands for HSI classification by the explainability of a convolutional neural network (CNN). We design a CNN architecture and use its 1D gradient-weighted class activation mapping (GradCAM) to obtain a gradient-weighted heatmap (GradHM) by the last layer in the well-trained CNN. Since the pooling layer in the CNN leads to a dimension change of the GradHM, cubic spline interpolation (CSI) is used to up-sample the GradHM. To further improve the accuracy of the up-sampled GradHM for the important band, guided backpropagation was adopted to obtain a more detailed GradHM (Guided-GradHM). Finally, based on GradHM and Guided-GradHM, two strategies, named Average Selection (AS) and Total Selection (TS), are proposed to form new and different combinations of methods (GradHM+AS, GradHM+TS, Guided-GradHM+AS, and Guided-GradHM+TS). In experiments, the proposed methods show better performance compared to other methods. In most cases, if the selected bands were fewer, then the Guided-GradHM+AS was more credible than the other methods. Furthermore, different datasets were used to validate the proposed methods. INDEX TERMS Explainability of CNN, hyperspectral, gradient-weighted class activation, guided backpropagation, band selection.

show abstract

“…Some algorithms have been reviewed in [10], where the hyperspectral bands are ranked by measures such as the Shannon entropy or spectral derivatives. Other approaches include band clustering 35 using various similarity measures and selecting representatives [11,12]. Popular similarity measures include information theoretical measures [13] or the correlation coefficient [14].…”

Section: Accepted Manuscriptmentioning

confidence: 99%

MIMR-DGSA: Unsupervised hyperspectral band selection based on information theory and a modified discrete gravitational search algorithm

et al. 2019

View full text Add to dashboard Cite

Band selection plays an important role in hyperspectral data analysis as it can improve the performance of data analysis without losing information about the constitution of the underlying data. We propose a MIMR-DGSA algorithm for band selection by following the Maximum-Information-Minimum-Redundancy (MIMR) criterion that maximises the information carried by individual features of a subset and minimises redundant information between them. Subsets are generated with a modified Discrete Gravitational Search Algorithm (DGSA) where we definine a neighbourhood concept for feature subsets. A fast algorithm for pairwise mutual information calculation that incorporates variable bandwidths of hyperspectral bands called VarBWFastMI is also developed. Classification results on three hyperspectral remote sensing datasets show that the proposed MIMR-DGSA performs similar to the original MIMR with Clonal Selection Algorithm (CSA) but is computationally more efficient and easier to

show abstract

Representative band selection for hyperspectral image classification

Cited by 51 publications

References 38 publications

Entropy-Based Convex Set Optimization for Spatial–Spectral Endmember Extraction From Hyperspectral Images

Entropy-Based Convex Set Optimization for Spatial–Spectral Endmember Extraction From Hyperspectral Images

Band Selection via Explanations From Convolutional Neural Networks

MIMR-DGSA: Unsupervised hyperspectral band selection based on information theory and a modified discrete gravitational search algorithm

Contact Info

Product

Resources

About