A Novel Image Classification Algorithm Using Overcomplete Wavelet Transforms

Soe Myint, Tong Zhu, Baojuan Zheng

Research output: Contribution to journalArticle

8 Citations (Scopus)

Abstract

A novel frequency-based classification framework and new wavelet algorithm (Wave-CLASS) is proposed using an overcomplete decomposition procedure. This approach omits the downsampling procedure and produces four-texture information with the same dimension of the original image or window at infinite scale. Three image subsets of QuickBird data (i.e., park, commercial, and rural) over a central region in the city of Phoenix were used to examine the effectiveness of the new wavelet overcomplete algorithm in comparison with a widely used classical approach (i.e., maximum likelihood). While the maximum-likelihood classifier produced < 78.29% overall accuracies for all three image subsets, the Wave-CLASS algorithm achieved high overall accuracies - 95.05% for the commercial subset (Kappa = 0.94), 93.71% for the park subset (Kappa = 0.93), and 89.33% for the rural subset (Kappa = 0.86). Results from this study demonstrate that the proposed method is effective in identifying detailed urban land cover types in high spatial resolution data.

Original languageEnglish (US)
Article number7047799
Pages (from-to)1232-1236
Number of pages5
JournalIEEE Geoscience and Remote Sensing Letters
Volume12
Issue number6
DOIs
StatePublished - 2015

Fingerprint

Image classification
image classification
Wavelet transforms
wavelet
transform
Maximum likelihood
QuickBird
Set theory
land cover
spatial resolution
Classifiers
Textures
texture
decomposition
Decomposition

Keywords

  • Classification
  • high spatial resolution
  • infinite scale
  • overcomplete decomposition
  • urban land cover
  • wavelet transforms

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Geotechnical Engineering and Engineering Geology

Cite this

A Novel Image Classification Algorithm Using Overcomplete Wavelet Transforms. / Myint, Soe; Zhu, Tong; Zheng, Baojuan.

In: IEEE Geoscience and Remote Sensing Letters, Vol. 12, No. 6, 7047799, 2015, p. 1232-1236.

Research output: Contribution to journalArticle

@article{1b7e4b9b45344576862d2b1733bd3487,
title = "A Novel Image Classification Algorithm Using Overcomplete Wavelet Transforms",
abstract = "A novel frequency-based classification framework and new wavelet algorithm (Wave-CLASS) is proposed using an overcomplete decomposition procedure. This approach omits the downsampling procedure and produces four-texture information with the same dimension of the original image or window at infinite scale. Three image subsets of QuickBird data (i.e., park, commercial, and rural) over a central region in the city of Phoenix were used to examine the effectiveness of the new wavelet overcomplete algorithm in comparison with a widely used classical approach (i.e., maximum likelihood). While the maximum-likelihood classifier produced < 78.29{\%} overall accuracies for all three image subsets, the Wave-CLASS algorithm achieved high overall accuracies - 95.05{\%} for the commercial subset (Kappa = 0.94), 93.71{\%} for the park subset (Kappa = 0.93), and 89.33{\%} for the rural subset (Kappa = 0.86). Results from this study demonstrate that the proposed method is effective in identifying detailed urban land cover types in high spatial resolution data.",
keywords = "Classification, high spatial resolution, infinite scale, overcomplete decomposition, urban land cover, wavelet transforms",
author = "Soe Myint and Tong Zhu and Baojuan Zheng",
year = "2015",
doi = "10.1109/LGRS.2015.2390133",
language = "English (US)",
volume = "12",
pages = "1232--1236",
journal = "IEEE Geoscience and Remote Sensing Letters",
issn = "1545-598X",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "6",

}

TY - JOUR

T1 - A Novel Image Classification Algorithm Using Overcomplete Wavelet Transforms

AU - Myint, Soe

AU - Zhu, Tong

AU - Zheng, Baojuan

PY - 2015

Y1 - 2015

N2 - A novel frequency-based classification framework and new wavelet algorithm (Wave-CLASS) is proposed using an overcomplete decomposition procedure. This approach omits the downsampling procedure and produces four-texture information with the same dimension of the original image or window at infinite scale. Three image subsets of QuickBird data (i.e., park, commercial, and rural) over a central region in the city of Phoenix were used to examine the effectiveness of the new wavelet overcomplete algorithm in comparison with a widely used classical approach (i.e., maximum likelihood). While the maximum-likelihood classifier produced < 78.29% overall accuracies for all three image subsets, the Wave-CLASS algorithm achieved high overall accuracies - 95.05% for the commercial subset (Kappa = 0.94), 93.71% for the park subset (Kappa = 0.93), and 89.33% for the rural subset (Kappa = 0.86). Results from this study demonstrate that the proposed method is effective in identifying detailed urban land cover types in high spatial resolution data.

AB - A novel frequency-based classification framework and new wavelet algorithm (Wave-CLASS) is proposed using an overcomplete decomposition procedure. This approach omits the downsampling procedure and produces four-texture information with the same dimension of the original image or window at infinite scale. Three image subsets of QuickBird data (i.e., park, commercial, and rural) over a central region in the city of Phoenix were used to examine the effectiveness of the new wavelet overcomplete algorithm in comparison with a widely used classical approach (i.e., maximum likelihood). While the maximum-likelihood classifier produced < 78.29% overall accuracies for all three image subsets, the Wave-CLASS algorithm achieved high overall accuracies - 95.05% for the commercial subset (Kappa = 0.94), 93.71% for the park subset (Kappa = 0.93), and 89.33% for the rural subset (Kappa = 0.86). Results from this study demonstrate that the proposed method is effective in identifying detailed urban land cover types in high spatial resolution data.

KW - Classification

KW - high spatial resolution

KW - infinite scale

KW - overcomplete decomposition

KW - urban land cover

KW - wavelet transforms

UR - http://www.scopus.com/inward/record.url?scp=84924942997&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84924942997&partnerID=8YFLogxK

U2 - 10.1109/LGRS.2015.2390133

DO - 10.1109/LGRS.2015.2390133

M3 - Article

VL - 12

SP - 1232

EP - 1236

JO - IEEE Geoscience and Remote Sensing Letters

JF - IEEE Geoscience and Remote Sensing Letters

SN - 1545-598X

IS - 6

M1 - 7047799

ER -