REGION-OF-INTEREST EXTRACTION IN LOW DEPTH OF FIELD IMAGES USING ENSEMBLE CLUSTERING AND DIFFERENCE OF GAUSSIAN APPROACHES

G. Rafiee, S.S. Dlay, W.L. Woo

Research output: Contribution to journalArticle

20 Citations (Scopus)

Abstract

In this paper, a two-stage unsupervised segmentation approach based on ensemble clustering is proposed to extract the focused regions from low depth-of-field (DOF) images. The first stage is to cluster image blocks in a joint contrast-energy feature space into three constituent groups. To achieve this, we make use of a normal mixture-based model along with standard expectation-maximization (EM) algorithm at two consecutive levels of block size. To avoid the common problem of local optima experienced in many models, an ensemble EM clustering algorithm is proposed. As a result, relevant blocks closely conforming to image objects are extracted. In stage two, a binary saliency map is constructed from the relevant blocks at the pixel level, which is based on difference of Gaussian (DOG) and binarization methods. Then, a set of morphological operations is employed to create the region-of-interest (ROI) from the map. Experimental results demonstrate that the proposed approach achieves an F-measure of 91.3% and is computationally 3 times faster than the existing state-of-the-art approach.
Original languageEnglish
Pages (from-to)2685-2699
Number of pages15
JournalPattern Recognition
Volume46
Issue number10
DOIs
Publication statusPublished - 2013

Fingerprint

Clustering algorithms
Pixels

Keywords

  • Low depth-of-field
  • Difference of Gaussian method
  • Ensemble clustering
  • Expectation-maximization algorithm
  • Region-of-interest extraction

Cite this

@article{90fc8b87390640ddbeff2c127fb4d7ab,
title = "REGION-OF-INTEREST EXTRACTION IN LOW DEPTH OF FIELD IMAGES USING ENSEMBLE CLUSTERING AND DIFFERENCE OF GAUSSIAN APPROACHES",
abstract = "In this paper, a two-stage unsupervised segmentation approach based on ensemble clustering is proposed to extract the focused regions from low depth-of-field (DOF) images. The first stage is to cluster image blocks in a joint contrast-energy feature space into three constituent groups. To achieve this, we make use of a normal mixture-based model along with standard expectation-maximization (EM) algorithm at two consecutive levels of block size. To avoid the common problem of local optima experienced in many models, an ensemble EM clustering algorithm is proposed. As a result, relevant blocks closely conforming to image objects are extracted. In stage two, a binary saliency map is constructed from the relevant blocks at the pixel level, which is based on difference of Gaussian (DOG) and binarization methods. Then, a set of morphological operations is employed to create the region-of-interest (ROI) from the map. Experimental results demonstrate that the proposed approach achieves an F-measure of 91.3{\%} and is computationally 3 times faster than the existing state-of-the-art approach.",
keywords = "Low depth-of-field, Difference of Gaussian method, Ensemble clustering, Expectation-maximization algorithm, Region-of-interest extraction",
author = "G. Rafiee and S.S. Dlay and W.L. Woo",
year = "2013",
doi = "10.1016/j.patcog.2013.03.006",
language = "English",
volume = "46",
pages = "2685--2699",
journal = "Pattern Recognition",
issn = "0031-3203",
publisher = "Elsevier Limited",
number = "10",

}

REGION-OF-INTEREST EXTRACTION IN LOW DEPTH OF FIELD IMAGES USING ENSEMBLE CLUSTERING AND DIFFERENCE OF GAUSSIAN APPROACHES. / Rafiee, G.; Dlay, S.S.; Woo, W.L.

In: Pattern Recognition, Vol. 46, No. 10, 2013, p. 2685-2699.

Research output: Contribution to journalArticle

TY - JOUR

T1 - REGION-OF-INTEREST EXTRACTION IN LOW DEPTH OF FIELD IMAGES USING ENSEMBLE CLUSTERING AND DIFFERENCE OF GAUSSIAN APPROACHES

AU - Rafiee, G.

AU - Dlay, S.S.

AU - Woo, W.L.

PY - 2013

Y1 - 2013

N2 - In this paper, a two-stage unsupervised segmentation approach based on ensemble clustering is proposed to extract the focused regions from low depth-of-field (DOF) images. The first stage is to cluster image blocks in a joint contrast-energy feature space into three constituent groups. To achieve this, we make use of a normal mixture-based model along with standard expectation-maximization (EM) algorithm at two consecutive levels of block size. To avoid the common problem of local optima experienced in many models, an ensemble EM clustering algorithm is proposed. As a result, relevant blocks closely conforming to image objects are extracted. In stage two, a binary saliency map is constructed from the relevant blocks at the pixel level, which is based on difference of Gaussian (DOG) and binarization methods. Then, a set of morphological operations is employed to create the region-of-interest (ROI) from the map. Experimental results demonstrate that the proposed approach achieves an F-measure of 91.3% and is computationally 3 times faster than the existing state-of-the-art approach.

AB - In this paper, a two-stage unsupervised segmentation approach based on ensemble clustering is proposed to extract the focused regions from low depth-of-field (DOF) images. The first stage is to cluster image blocks in a joint contrast-energy feature space into three constituent groups. To achieve this, we make use of a normal mixture-based model along with standard expectation-maximization (EM) algorithm at two consecutive levels of block size. To avoid the common problem of local optima experienced in many models, an ensemble EM clustering algorithm is proposed. As a result, relevant blocks closely conforming to image objects are extracted. In stage two, a binary saliency map is constructed from the relevant blocks at the pixel level, which is based on difference of Gaussian (DOG) and binarization methods. Then, a set of morphological operations is employed to create the region-of-interest (ROI) from the map. Experimental results demonstrate that the proposed approach achieves an F-measure of 91.3% and is computationally 3 times faster than the existing state-of-the-art approach.

KW - Low depth-of-field

KW - Difference of Gaussian method

KW - Ensemble clustering

KW - Expectation-maximization algorithm

KW - Region-of-interest extraction

U2 - 10.1016/j.patcog.2013.03.006

DO - 10.1016/j.patcog.2013.03.006

M3 - Article

VL - 46

SP - 2685

EP - 2699

JO - Pattern Recognition

JF - Pattern Recognition

SN - 0031-3203

IS - 10

ER -