Non-parametric human segmentation using support vector machine

Kyuwon Kim, Changjae Oh, Kwanghoon Sohn

Research output: Contribution to journalArticle

7 Citations (Scopus)

Abstract

The human category is among the most commonly captured and therefore important subject in digital cameras. As a result, human segmentation also plays an important role in the camera system. This paper presents a framework of non-parametric learning that achieves human segmentation using support vector machine (SVM). In the training stage, the training human windows are warped to a normalized size and oversegmented into regular superpixels. A two-dimensional (2D) array of SVM models is then trained by extracting various edge and color features from each superpixel. In this process of SVM training, the 2D array of SVM models automatically and effectively learns the characteristics of the human shape. Given an affinely warped human window for testing, the proposed method calculates superpixels' initial scores of belonging to a human (foreground) using the trained SVM models. Finally, the initial prediction scores are effectively propagated by optimizing a well-defined energy function using an estimated confidence map. In experiments on a publicly available challenging dataset, the proposed framework rapidly yields excellent results in human segmentation both qualitatively and quantitatively1.

Original languageEnglish
Article number7514714
Pages (from-to)150-158
Number of pages9
JournalIEEE Transactions on Consumer Electronics
Volume62
Issue number2
DOIs
Publication statusPublished - 2016 May 1

Fingerprint

Support vector machines
Digital cameras
Cameras
Color
Testing
Experiments

All Science Journal Classification (ASJC) codes

  • Media Technology
  • Electrical and Electronic Engineering

Cite this

@article{75b21964264d47fbb0a3e8968b503d2b,
title = "Non-parametric human segmentation using support vector machine",
abstract = "The human category is among the most commonly captured and therefore important subject in digital cameras. As a result, human segmentation also plays an important role in the camera system. This paper presents a framework of non-parametric learning that achieves human segmentation using support vector machine (SVM). In the training stage, the training human windows are warped to a normalized size and oversegmented into regular superpixels. A two-dimensional (2D) array of SVM models is then trained by extracting various edge and color features from each superpixel. In this process of SVM training, the 2D array of SVM models automatically and effectively learns the characteristics of the human shape. Given an affinely warped human window for testing, the proposed method calculates superpixels' initial scores of belonging to a human (foreground) using the trained SVM models. Finally, the initial prediction scores are effectively propagated by optimizing a well-defined energy function using an estimated confidence map. In experiments on a publicly available challenging dataset, the proposed framework rapidly yields excellent results in human segmentation both qualitatively and quantitatively1.",
author = "Kyuwon Kim and Changjae Oh and Kwanghoon Sohn",
year = "2016",
month = "5",
day = "1",
doi = "10.1109/TCE.2016.7514714",
language = "English",
volume = "62",
pages = "150--158",
journal = "IEEE Transactions on Consumer Electronics",
issn = "0098-3063",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "2",

}

Non-parametric human segmentation using support vector machine. / Kim, Kyuwon; Oh, Changjae; Sohn, Kwanghoon.

In: IEEE Transactions on Consumer Electronics, Vol. 62, No. 2, 7514714, 01.05.2016, p. 150-158.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Non-parametric human segmentation using support vector machine

AU - Kim, Kyuwon

AU - Oh, Changjae

AU - Sohn, Kwanghoon

PY - 2016/5/1

Y1 - 2016/5/1

N2 - The human category is among the most commonly captured and therefore important subject in digital cameras. As a result, human segmentation also plays an important role in the camera system. This paper presents a framework of non-parametric learning that achieves human segmentation using support vector machine (SVM). In the training stage, the training human windows are warped to a normalized size and oversegmented into regular superpixels. A two-dimensional (2D) array of SVM models is then trained by extracting various edge and color features from each superpixel. In this process of SVM training, the 2D array of SVM models automatically and effectively learns the characteristics of the human shape. Given an affinely warped human window for testing, the proposed method calculates superpixels' initial scores of belonging to a human (foreground) using the trained SVM models. Finally, the initial prediction scores are effectively propagated by optimizing a well-defined energy function using an estimated confidence map. In experiments on a publicly available challenging dataset, the proposed framework rapidly yields excellent results in human segmentation both qualitatively and quantitatively1.

AB - The human category is among the most commonly captured and therefore important subject in digital cameras. As a result, human segmentation also plays an important role in the camera system. This paper presents a framework of non-parametric learning that achieves human segmentation using support vector machine (SVM). In the training stage, the training human windows are warped to a normalized size and oversegmented into regular superpixels. A two-dimensional (2D) array of SVM models is then trained by extracting various edge and color features from each superpixel. In this process of SVM training, the 2D array of SVM models automatically and effectively learns the characteristics of the human shape. Given an affinely warped human window for testing, the proposed method calculates superpixels' initial scores of belonging to a human (foreground) using the trained SVM models. Finally, the initial prediction scores are effectively propagated by optimizing a well-defined energy function using an estimated confidence map. In experiments on a publicly available challenging dataset, the proposed framework rapidly yields excellent results in human segmentation both qualitatively and quantitatively1.

UR - http://www.scopus.com/inward/record.url?scp=84979516762&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84979516762&partnerID=8YFLogxK

U2 - 10.1109/TCE.2016.7514714

DO - 10.1109/TCE.2016.7514714

M3 - Article

VL - 62

SP - 150

EP - 158

JO - IEEE Transactions on Consumer Electronics

JF - IEEE Transactions on Consumer Electronics

SN - 0098-3063

IS - 2

M1 - 7514714

ER -