Tampere University of Technology

TUTCRIS Research Portal

Learning to rank salient segments extracted by multispectral quantum cuts

Research output: Contribution to journalArticleScientificpeer-review

Standard

Learning to rank salient segments extracted by multispectral quantum cuts. / Aytekin, Caglar; Kiranyaz, Serkan; Gabbouj, Moncef.

In: Pattern Recognition Letters, Vol. 72, 03.2016, p. 91-99.

Research output: Contribution to journalArticleScientificpeer-review

Harvard

APA

Vancouver

Author

Aytekin, Caglar ; Kiranyaz, Serkan ; Gabbouj, Moncef. / Learning to rank salient segments extracted by multispectral quantum cuts. In: Pattern Recognition Letters. 2016 ; Vol. 72. pp. 91-99.

Bibtex - Download

@article{75449e5eee7e4e9e9881c199ff1b1d4f,
title = "Learning to rank salient segments extracted by multispectral quantum cuts",
abstract = "In this paper, a learn-to-rank algorithm is proposed and applied over the segment pool of salient objects generated by an extension of the unsupervised Quantum-Cuts algorithm. The existing Quantum Cuts is extended in a multiresolution approach as follows. First, superpixels are extracted from the input image using the simple linear iterative k-means algorithm; second, a scale space decomposition is applied prior to Quantum Cuts in order to capture salient details at different scales; and third, multispectral approach is followed to generate multiple proposals instead of a single proposal as in Quantum Cuts. The proposed learn-to-rank algorithm is then applied to these multiple proposals in order to select the most appropriate one. Shape and appearance features are extracted from the proposed segments and regressed with respect to a given confidence measure resulting in a ranked list of proposals. This ranking yields consistent improvements in an extensive collection of benchmark datasets containing around 18k images. Our analysis on the random forest regression models that are trained on different datasets shows that, although these datasets are of quite different characteristics, a model trained in the most complex dataset consistently provides performance improvements in all the other datasets, hence yielding robust salient object segmentation with a significant performance gap compared to the competing methods.",
keywords = "Quantum Cuts, Saliency Detection, Learning to Rank, Salient Object Segmentation, Multispectral Analysis",
author = "Caglar Aytekin and Serkan Kiranyaz and Moncef Gabbouj",
note = "http://www.sciencedirect.com/science/article/pii/S0167865515004286",
year = "2016",
month = "3",
doi = "10.1016/j.patrec.2015.12.005",
language = "English",
volume = "72",
pages = "91--99",
journal = "Pattern Recognition Letters",
issn = "0167-8655",
publisher = "Elsevier",

}

RIS (suitable for import to EndNote) - Download

TY - JOUR

T1 - Learning to rank salient segments extracted by multispectral quantum cuts

AU - Aytekin, Caglar

AU - Kiranyaz, Serkan

AU - Gabbouj, Moncef

N1 - http://www.sciencedirect.com/science/article/pii/S0167865515004286

PY - 2016/3

Y1 - 2016/3

N2 - In this paper, a learn-to-rank algorithm is proposed and applied over the segment pool of salient objects generated by an extension of the unsupervised Quantum-Cuts algorithm. The existing Quantum Cuts is extended in a multiresolution approach as follows. First, superpixels are extracted from the input image using the simple linear iterative k-means algorithm; second, a scale space decomposition is applied prior to Quantum Cuts in order to capture salient details at different scales; and third, multispectral approach is followed to generate multiple proposals instead of a single proposal as in Quantum Cuts. The proposed learn-to-rank algorithm is then applied to these multiple proposals in order to select the most appropriate one. Shape and appearance features are extracted from the proposed segments and regressed with respect to a given confidence measure resulting in a ranked list of proposals. This ranking yields consistent improvements in an extensive collection of benchmark datasets containing around 18k images. Our analysis on the random forest regression models that are trained on different datasets shows that, although these datasets are of quite different characteristics, a model trained in the most complex dataset consistently provides performance improvements in all the other datasets, hence yielding robust salient object segmentation with a significant performance gap compared to the competing methods.

AB - In this paper, a learn-to-rank algorithm is proposed and applied over the segment pool of salient objects generated by an extension of the unsupervised Quantum-Cuts algorithm. The existing Quantum Cuts is extended in a multiresolution approach as follows. First, superpixels are extracted from the input image using the simple linear iterative k-means algorithm; second, a scale space decomposition is applied prior to Quantum Cuts in order to capture salient details at different scales; and third, multispectral approach is followed to generate multiple proposals instead of a single proposal as in Quantum Cuts. The proposed learn-to-rank algorithm is then applied to these multiple proposals in order to select the most appropriate one. Shape and appearance features are extracted from the proposed segments and regressed with respect to a given confidence measure resulting in a ranked list of proposals. This ranking yields consistent improvements in an extensive collection of benchmark datasets containing around 18k images. Our analysis on the random forest regression models that are trained on different datasets shows that, although these datasets are of quite different characteristics, a model trained in the most complex dataset consistently provides performance improvements in all the other datasets, hence yielding robust salient object segmentation with a significant performance gap compared to the competing methods.

KW - Quantum Cuts

KW - Saliency Detection

KW - Learning to Rank

KW - Salient Object Segmentation

KW - Multispectral Analysis

U2 - 10.1016/j.patrec.2015.12.005

DO - 10.1016/j.patrec.2015.12.005

M3 - Article

VL - 72

SP - 91

EP - 99

JO - Pattern Recognition Letters

JF - Pattern Recognition Letters

SN - 0167-8655

ER -