2020
DOI: 10.3390/rs12182976
|View full text |Cite
|
Sign up to set email alerts
|

Shape Adaptive Neighborhood Information-Based Semi-Supervised Learning for Hyperspectral Image Classification

Abstract: Hyperspectral image (HSI) classification is an important research topic in detailed analysis of the Earth’s surface. However, the performance of the classification is often hampered by the high-dimensionality features and limited training samples of the HSIs which has fostered research about semi-supervised learning (SSL). In this paper, we propose a shape adaptive neighborhood information (SANI) based SSL (SANI-SSL) method that takes full advantage of the adaptive spatial information to select valuable unlabe… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
3
2

Citation Types

0
8
0

Year Published

2021
2021
2024
2024

Publication Types

Select...
7
1

Relationship

0
8

Authors

Journals

citations
Cited by 10 publications
(8 citation statements)
references
References 49 publications
0
8
0
Order By: Relevance
“…Spatial lags ∆1 and ∆2 along each axis of coordinate are generally assumed to be 1. According to (6), we calculate the noise covariance Σ N .…”
Section: Feature Extractionmentioning
confidence: 99%
See 2 more Smart Citations
“…Spatial lags ∆1 and ∆2 along each axis of coordinate are generally assumed to be 1. According to (6), we calculate the noise covariance Σ N .…”
Section: Feature Extractionmentioning
confidence: 99%
“…Σ N = 0.5 (Noise image) (6) Filters can be applied to the bandwidth that has the least noise level; this effectively removes noise from the data. Then, the data is transmitted back to its original coordinate system.…”
Section: Feature Extractionmentioning
confidence: 99%
See 1 more Smart Citation
“…Some authors have implemented neighbor selection using threshold, for example, in hyperspectral unmixing applications [ 18 ]. Others have implemented adaptive shapes [ 19 ] in classification methods that does not involve GCNs. In this work, we propose a novel way of adaptively creating the adjacency matrix based on a neighbor selection approach called AN-GCN.…”
Section: Introductionmentioning
confidence: 99%
“…As methods for the feature extraction of hyperspectral data are being continuously improved, traditional classification methods, based on spectral features alone [8,9], no longer meet requirements for high classification accuracy. As a result, numerous scholars have successfully incorporated spatial information into classification and have proven the ability to improve the classification performance [10][11][12]. However, spatial-spectral features are commonly used as classification in only two ways: first, one can extract spectral and spatial features separately and input their stacked joint features into the classifier for classification; second, joint spatial-spectral features can be extracted simultaneously for classification.…”
Section: Introductionmentioning
confidence: 99%