IDA: Improving distribution analysis for reducing data complexity and dimensionality in hyperspectral images

被引：9

作者：

AL-Alimi, Dalal ^{[1
,3
]}

Al-qaness, Mohammed A. ^{[2
,3
,4
]}

Cai, Zhihua ^{[1
]}

Alawamy, Eman Ahmed ^{[5
]}

机构：

[1] China Univ Geosci, Sch Comp Sci, Wuhan 430074, Peoples R China

[2] Zhejiang Normal Univ, Coll Phys & Elect Informat Engn, Jinhua 321004, Peoples R China

[3] Sanaa Univ, Fac Engn, Sanaa 12544, Yemen

[4] Wuhan Univ, State Key Lab Informat Engn Surveying Mapping & Re, Wuhan 430079, Peoples R China

[5] Cent South Univ, Sch Math & Stat, Changsha 410083, Peoples R China

来源：

PATTERN RECOGNITION | 2023年 / 134卷

关键词：

Feature reduction; Hyperspectral image; Classification; Feature fusion; Feature extraction; Dimensionality reduction; CLASSIFICATION; NETWORK; FUSION; CNN;

D O I：

10.1016/j.patcog.2022.109096

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

Hyperspectral images (HSIs) are known for their high dimensionality and wide spectral bands that in-crease redundant information and complicate classification. Outliers and mixed data are common prob-lems in HSIs. Thus, preprocessing methods are essential in enhancing and reducing data complexity, re-dundant information, and the number of bands. This study introduces a novel feature reduction method (FRM) called improving distribution analysis (IDA). IDA works to increase the correlation between related data, decrease the distance between big and small data, and correct each value's location to be inside its group range. In IDA, the input data passes through three stages. Getting rid of outliers and improv-ing data correlation is the first step. The second stage involves increasing the variance. The third is to simplify the data and normalize the distribution. IDA is compared with four popular FRMs in four avail-able HSIs. It is also tested and evaluated in various classification models, including spatial, spectral, and spectral-spatial models. The experimental results demonstrate that IDA performs admirably in enhancing data distribution, reducing complexity, and accelerating performance.(c) 2022 Elsevier Ltd. All rights reserved.

引用

页数：18

共 40 条

[1] Applications of singular-value decomposition (SVD) [J].

Akritas, AG ;

Malaschonok, GI .

MATHEMATICS AND COMPUTERS IN SIMULATION, 2004, 67 (1-2) :15-31

[2] Conceptual and empirical comparison of dimensionality reduction algorithms (PCA, KPCA, LDA, MDS, SVD, LLE, ISOMAP, LE, ICA, t-SNE) [J].

Anowar, Farzana ;

Sadaoui, Samira ;

Selim, Bassant .

COMPUTER SCIENCE REVIEW, 2021, 40

[3] Feature selection in machine learning: A new perspective [J].

Cai, Jie ;

Luo, Jiawei ;

Wang, Shulin ;

Yang, Sheng .

NEUROCOMPUTING, 2018, 300 :70-79

[4] Recent Advances in Supervised Dimension Reduction: A Survey [J].

Chao, Guoqing ;

Luo, Yuan ;

Ding, Weiping .

MACHINE LEARNING AND KNOWLEDGE EXTRACTION, 2019, 1 (01) :341-358

[5] Spectral-spatial feature fusion via dual-stream deep architecture for hyperspectral image classification [J].

Chen, Rong ;

Li, Guanghui .

INFRARED PHYSICS & TECHNOLOGY, 2021, 119

[6]

Christovam L.E., 2019, INT ARCH PHOTOGRAMM

[7] Multidimensional relation learning for hyperspectral image classification [J].

Fang, Jie ;

Cao, Xiaoqian .

NEUROCOMPUTING, 2020, 410 :211-219

[8] Hyperspectral Image Classification Method Based on 2D-3D CNN and Multibranch Feature Fusion [J].

Ge, Zixian ;

Cao, Guo ;

Li, Xuesong ;

Fu, Peng .

IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, 2020, 13 :5776-5788

[9]

Ghojogh B, 2019, Arxiv, DOI [arXiv:1905.02845, 10.48550/arXiv.1905.02845, DOI 10.48550/ARXIV.1905.02845]

[10]

Goodfellow I, 2016, ADAPT COMPUT MACH LE, P1

← 1 2 3 4 →