Context-aware encoding for clothing parsing

C. H. Yoo, Y. G. Shin, S. W. Kim, Sung-Jea Ko

Research output: Contribution to journalArticle

Abstract

Clothing parsing is a special type of semantic segmentation in which each pixel is assigned with clothing labels. Unlike general scene semantic segmentation, stylish match (e.g. skirts + blouse, jeans + T-shirt) is an important cue for recognising fine-grained categories in clothing parsing. In this Letter, the authors propose a context-aware outfit encoder (COE), as a side branch, that drives the convolutional neural network to take the stylish match into account for clothing parsing. The proposed COE provides information on matching clothes that can be utilised to improve the prediction accuracy of the base network significantly. Experimental results show that fully convolutional network and MobileNet with the COE improve the mean intersection of the union of those without the COE by 2.5 and 2.8%, respectively, on CFPD dataset.

Original languageEnglish
Pages (from-to)692-693
Number of pages2
JournalElectronics Letters
Volume55
Issue number12
DOIs
Publication statusPublished - 2019 Jun 13

Fingerprint

Semantics
Labels
Pixels
Neural networks

ASJC Scopus subject areas

  • Electrical and Electronic Engineering

Cite this

Context-aware encoding for clothing parsing. / Yoo, C. H.; Shin, Y. G.; Kim, S. W.; Ko, Sung-Jea.

In: Electronics Letters, Vol. 55, No. 12, 13.06.2019, p. 692-693.

Research output: Contribution to journalArticle

Yoo, C. H. ; Shin, Y. G. ; Kim, S. W. ; Ko, Sung-Jea. / Context-aware encoding for clothing parsing. In: Electronics Letters. 2019 ; Vol. 55, No. 12. pp. 692-693.
@article{a61ec140fde7426d94ad08fa0b4f72f7,
title = "Context-aware encoding for clothing parsing",
abstract = "Clothing parsing is a special type of semantic segmentation in which each pixel is assigned with clothing labels. Unlike general scene semantic segmentation, stylish match (e.g. skirts + blouse, jeans + T-shirt) is an important cue for recognising fine-grained categories in clothing parsing. In this Letter, the authors propose a context-aware outfit encoder (COE), as a side branch, that drives the convolutional neural network to take the stylish match into account for clothing parsing. The proposed COE provides information on matching clothes that can be utilised to improve the prediction accuracy of the base network significantly. Experimental results show that fully convolutional network and MobileNet with the COE improve the mean intersection of the union of those without the COE by 2.5 and 2.8{\%}, respectively, on CFPD dataset.",
author = "Yoo, {C. H.} and Shin, {Y. G.} and Kim, {S. W.} and Sung-Jea Ko",
year = "2019",
month = "6",
day = "13",
doi = "10.1049/el.2019.1213",
language = "English",
volume = "55",
pages = "692--693",
journal = "Electronics Letters",
issn = "0013-5194",
publisher = "Institution of Engineering and Technology",
number = "12",

}

TY - JOUR

T1 - Context-aware encoding for clothing parsing

AU - Yoo, C. H.

AU - Shin, Y. G.

AU - Kim, S. W.

AU - Ko, Sung-Jea

PY - 2019/6/13

Y1 - 2019/6/13

N2 - Clothing parsing is a special type of semantic segmentation in which each pixel is assigned with clothing labels. Unlike general scene semantic segmentation, stylish match (e.g. skirts + blouse, jeans + T-shirt) is an important cue for recognising fine-grained categories in clothing parsing. In this Letter, the authors propose a context-aware outfit encoder (COE), as a side branch, that drives the convolutional neural network to take the stylish match into account for clothing parsing. The proposed COE provides information on matching clothes that can be utilised to improve the prediction accuracy of the base network significantly. Experimental results show that fully convolutional network and MobileNet with the COE improve the mean intersection of the union of those without the COE by 2.5 and 2.8%, respectively, on CFPD dataset.

AB - Clothing parsing is a special type of semantic segmentation in which each pixel is assigned with clothing labels. Unlike general scene semantic segmentation, stylish match (e.g. skirts + blouse, jeans + T-shirt) is an important cue for recognising fine-grained categories in clothing parsing. In this Letter, the authors propose a context-aware outfit encoder (COE), as a side branch, that drives the convolutional neural network to take the stylish match into account for clothing parsing. The proposed COE provides information on matching clothes that can be utilised to improve the prediction accuracy of the base network significantly. Experimental results show that fully convolutional network and MobileNet with the COE improve the mean intersection of the union of those without the COE by 2.5 and 2.8%, respectively, on CFPD dataset.

UR - http://www.scopus.com/inward/record.url?scp=85067829447&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85067829447&partnerID=8YFLogxK

U2 - 10.1049/el.2019.1213

DO - 10.1049/el.2019.1213

M3 - Article

AN - SCOPUS:85067829447

VL - 55

SP - 692

EP - 693

JO - Electronics Letters

JF - Electronics Letters

SN - 0013-5194

IS - 12

ER -