What Makes a Style

Experimental Analysis of Fashion Prediction

Moeko Takagi, Edgar Simo Serra, Satoshi Iizuka, Hiroshi Ishikawa

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Citations (Scopus)

Abstract

In this work, we perform an experimental analysis of the differences of both how humans and machines see and distinguish fashion styles. For this purpose, we propose an expert-curated new dataset for fashion style prediction, which consists of 14 different fashion styles each with roughly 1,000 images of worn outfits. The dataset, with a total of 13,126 images, captures the diversity and complexity of modern fashion styles. We perform an extensive analysis of the dataset by benchmarking a wide variety of modern classification networks, and also perform an in-depth user study with both fashion-savvy and fashion-naïve users. Our results indicate that, although classification networks are able to outperform naive users, they are still far from the performance of savvy users, for which it is important to not only consider texture and color, but subtle differences in the combination of garments.

Original languageEnglish
Title of host publicationProceedings - 2017 IEEE International Conference on Computer Vision Workshops, ICCVW 2017
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages2247-2253
Number of pages7
Volume2018-January
ISBN (Electronic)9781538610343
DOIs
Publication statusPublished - 2018 Jan 19
Event16th IEEE International Conference on Computer Vision Workshops, ICCVW 2017 - Venice, Italy
Duration: 2017 Oct 222017 Oct 29

Other

Other16th IEEE International Conference on Computer Vision Workshops, ICCVW 2017
CountryItaly
CityVenice
Period17/10/2217/10/29

Fingerprint

Benchmarking
Textures
Color

ASJC Scopus subject areas

  • Computer Science Applications
  • Computer Vision and Pattern Recognition

Cite this

Takagi, M., Simo Serra, E., Iizuka, S., & Ishikawa, H. (2018). What Makes a Style: Experimental Analysis of Fashion Prediction. In Proceedings - 2017 IEEE International Conference on Computer Vision Workshops, ICCVW 2017 (Vol. 2018-January, pp. 2247-2253). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICCVW.2017.263

What Makes a Style : Experimental Analysis of Fashion Prediction. / Takagi, Moeko; Simo Serra, Edgar; Iizuka, Satoshi; Ishikawa, Hiroshi.

Proceedings - 2017 IEEE International Conference on Computer Vision Workshops, ICCVW 2017. Vol. 2018-January Institute of Electrical and Electronics Engineers Inc., 2018. p. 2247-2253.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Takagi, M, Simo Serra, E, Iizuka, S & Ishikawa, H 2018, What Makes a Style: Experimental Analysis of Fashion Prediction. in Proceedings - 2017 IEEE International Conference on Computer Vision Workshops, ICCVW 2017. vol. 2018-January, Institute of Electrical and Electronics Engineers Inc., pp. 2247-2253, 16th IEEE International Conference on Computer Vision Workshops, ICCVW 2017, Venice, Italy, 17/10/22. https://doi.org/10.1109/ICCVW.2017.263
Takagi M, Simo Serra E, Iizuka S, Ishikawa H. What Makes a Style: Experimental Analysis of Fashion Prediction. In Proceedings - 2017 IEEE International Conference on Computer Vision Workshops, ICCVW 2017. Vol. 2018-January. Institute of Electrical and Electronics Engineers Inc. 2018. p. 2247-2253 https://doi.org/10.1109/ICCVW.2017.263
Takagi, Moeko ; Simo Serra, Edgar ; Iizuka, Satoshi ; Ishikawa, Hiroshi. / What Makes a Style : Experimental Analysis of Fashion Prediction. Proceedings - 2017 IEEE International Conference on Computer Vision Workshops, ICCVW 2017. Vol. 2018-January Institute of Electrical and Electronics Engineers Inc., 2018. pp. 2247-2253
@inproceedings{81fcf510673e496c8c5fbd300eea4f05,
title = "What Makes a Style: Experimental Analysis of Fashion Prediction",
abstract = "In this work, we perform an experimental analysis of the differences of both how humans and machines see and distinguish fashion styles. For this purpose, we propose an expert-curated new dataset for fashion style prediction, which consists of 14 different fashion styles each with roughly 1,000 images of worn outfits. The dataset, with a total of 13,126 images, captures the diversity and complexity of modern fashion styles. We perform an extensive analysis of the dataset by benchmarking a wide variety of modern classification networks, and also perform an in-depth user study with both fashion-savvy and fashion-na{\"i}ve users. Our results indicate that, although classification networks are able to outperform naive users, they are still far from the performance of savvy users, for which it is important to not only consider texture and color, but subtle differences in the combination of garments.",
author = "Moeko Takagi and {Simo Serra}, Edgar and Satoshi Iizuka and Hiroshi Ishikawa",
year = "2018",
month = "1",
day = "19",
doi = "10.1109/ICCVW.2017.263",
language = "English",
volume = "2018-January",
pages = "2247--2253",
booktitle = "Proceedings - 2017 IEEE International Conference on Computer Vision Workshops, ICCVW 2017",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - What Makes a Style

T2 - Experimental Analysis of Fashion Prediction

AU - Takagi, Moeko

AU - Simo Serra, Edgar

AU - Iizuka, Satoshi

AU - Ishikawa, Hiroshi

PY - 2018/1/19

Y1 - 2018/1/19

N2 - In this work, we perform an experimental analysis of the differences of both how humans and machines see and distinguish fashion styles. For this purpose, we propose an expert-curated new dataset for fashion style prediction, which consists of 14 different fashion styles each with roughly 1,000 images of worn outfits. The dataset, with a total of 13,126 images, captures the diversity and complexity of modern fashion styles. We perform an extensive analysis of the dataset by benchmarking a wide variety of modern classification networks, and also perform an in-depth user study with both fashion-savvy and fashion-naïve users. Our results indicate that, although classification networks are able to outperform naive users, they are still far from the performance of savvy users, for which it is important to not only consider texture and color, but subtle differences in the combination of garments.

AB - In this work, we perform an experimental analysis of the differences of both how humans and machines see and distinguish fashion styles. For this purpose, we propose an expert-curated new dataset for fashion style prediction, which consists of 14 different fashion styles each with roughly 1,000 images of worn outfits. The dataset, with a total of 13,126 images, captures the diversity and complexity of modern fashion styles. We perform an extensive analysis of the dataset by benchmarking a wide variety of modern classification networks, and also perform an in-depth user study with both fashion-savvy and fashion-naïve users. Our results indicate that, although classification networks are able to outperform naive users, they are still far from the performance of savvy users, for which it is important to not only consider texture and color, but subtle differences in the combination of garments.

UR - http://www.scopus.com/inward/record.url?scp=85046291667&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85046291667&partnerID=8YFLogxK

U2 - 10.1109/ICCVW.2017.263

DO - 10.1109/ICCVW.2017.263

M3 - Conference contribution

VL - 2018-January

SP - 2247

EP - 2253

BT - Proceedings - 2017 IEEE International Conference on Computer Vision Workshops, ICCVW 2017

PB - Institute of Electrical and Electronics Engineers Inc.

ER -