Home / new style clothing for women / What dress fits me best? Fashion Recommendation …
What dress fits me best? Fashion Recommendation on the
Clothing Style for Personal Body Shape
Shintami Chusnul Hidayati Cheng-Chun Hsu Yu-Ting Chang
Research Center for IT Innovation Dept. of CSIE Research Center for IT Innovation
Academia Sinica Nat'l Taiwan Univ. of Sci. and Tech. Academia Sinica
Taipei, Taiwan Taipei, Taiwan Taipei, Taiwan
shintami@citi.sinica.edu.tw b10415009@mail.ntust.edu.tw juliachang@citi.sinica.edu.tw
Kai-Lung Hua Jianlong Fu Wen-Huang Cheng
Dept. of CSIE Microsoft Research CITI, Academia Sinica
Nat'l Taiwan Univ. of Sci. and Tech. Beijing, China EE, National Chiao Tung University
Taipei, Taiwan jianf@microsoft.com Taipei/Hsinchu, Taiwan
hua@mail.ntust.edu.tw whcheng@nctu.edu.tw
ABSTRACT
Clothing is an integral part of life. Also, it is always an uneasy task
for people to make decisions on what to wear. An essential style
tip is to dress for the body shape, i.e., knowing one's own body
shape (e.g., hourglass, rectangle, round and inverted triangle) and
selecting the types of clothes that will accentuate the body's good
features. In the literature, although various fashion recommenda-
tion systems for clothing items have been developed, none of them
had explicitly taken the user's basic body shape into consideration.
In this paper, therefore, we proposed a first framework for learning
the compatibility of clothing styles and body shapes from social big
data, with the goal to recommend a user about what to wear better
in relation to his/her essential body attributes. The experimental re-
sults demonstrate the superiority of our proposed approach, leading Figure 1: Visual illustration of five different body shapes.
to a new aspect for research into fashion recommendation. (Image courtesy of styleangel.com1)
CCS CONCEPTS
? Information systems Personalization; Recommender sys- 1 INTRODUCTION
tems; Information extraction; "The dress must follow the body of a woman, not
the body following the shape of the dress."
KEYWORDS -- Hubert de Givenchy, fashion designer
Fashion analysis; recommender system; human body shape; cloth- Nobody is perfect. Everyone has flaws and strengths, without
ing style; correlation exception to their body parts. Accordingly, everyone may have
ACM Reference Format: areas on their body they love to highlight and areas they feel like
Shintami Chusnul Hidayati, Cheng-Chun Hsu, Yu-Ting Chang, Kai-Lung covering up [1]. By understanding the type of body shapes and
Hua, Jianlong Fu, and Wen-Huang Cheng. 2018. What dress fits me best? what styles flatter and complement particular body shape, everyone
Fashion Recommendation on the Clothing Style for Personal Body Shape. surely can work with their body to give off amazing looks.
In 2018 ACM Multimedia Conference (MM '18), October 22-26, 2018, Seoul, Re- To the above challenge, body shape is the first thing to consider
public of Korea. ACM, New York, NY, USA, 9 pages. https://doi.org/10.1145/ before choosing the perfect clothing styles. In particular, determin-
3240508.3240546 ing the type of body shape is all about the proportions between
Permission to make digital or hard copies of all or part of this work for personal or body measurements. There are several methods to determine body
classroom use is granted without fee provided that copies are not made or distributed shapes, which are known as body shape calculators. They use differ-
for profit or commercial advantage and that copies bear this notice and the full citation ent representations to describe the type of body shapes, such as in
on the first page. Copyrights for components of this work owned by others than the
author(s) must be honored. Abstracting with credit is permitted. To copy otherwise, or terms of alphabetical letters (e.g., the S-line for a body shape with
republish, to post on servers or to redistribute to lists, requires prior specific permission ample breasts and buttocks when viewed from the side), fruit shapes
and/or a fee. Request permissions from permissions@acm.org. (e.g., the apple for a body shape with a wide torso, broad shoulders,
MM '18, October 22-26, 2018, Seoul, Republic of Korea
? 2018 Copyright held by the owner/author(s). Publication rights licensed to ACM. and a full bust, waist, and upper back), and geometric objects (e.g.,
ACM ISBN 978-1-4503-5665-7/18/10. . . $15.00
https://doi.org/10.1145/3240508.3240546 1https://styleangel.com/discover-your-body-shape/
reasonable statistical model that can capture the correlations be-
tween body shapes and clothing styles well. Therefore, we exploit
two kinds of data, i.e., human body measurements and clothing
styles. In particular, following the styles of our favorite stylish peo-
ple, regardless of whether they have the same body shape as ours,
will cause a low chance of these styles to suit our body shape; while
following the styles of people that we consider having similar body
shapes to ours requires an understanding of whether their styles
are fashionable or not. By incorporating both clothing style and
body shape information, we mine the auxiliary clothing features
to discover semantically important styles for each body shape. For
Figure 2: An example of guideline for choosing wedding this purpose, we construct graphs of images by visual and body
dress. Finding the right wedding dress is probably one of shape information, respectively. Then, we automatically propagate
the most crucial things for any bride to be. To find a gown the semantic and select the relevant styles across the visual and
with the perfect fit, it is essential to choose one which suits body shape graphs.
body shape and flatters the curves. (Image courtesy of vira- We summarize our main contributions as follows:
lyfeeds.com2) ? To the best of our knowledge, this is the first study in learning
the golden style rules to how to best flatter each body type
by exploiting big social data.
the top hourglass for a body shape with well defined waist and ? To obtain the relevant fashion knowledge rules, we present
larger bust than hips). Most common are the five basic body shapes, a mechanism of style propagation for discovering semantic
as shown in Figure 1. The existing body shape calculators, however, relations between clothing styles by considering body figures
are based on subjective measures, which are susceptible to multiple of people wearing it.
biases. To our best knowledge, there is no standardized method for ? We construct a body-style map to model the correlation be-
determining the type of body shapes. Therefore, determining the tween clothing styles and body shapes. Based on the body-
type of body shapes is still a challenging analytical task. style map, we can describe which clothing styles that suit
After knowing the type of body shape, we can find out what a particular body appearance, and vice versa, to provide a
kind of style fits best. The style of clothes affects appearance greatly personalized style suggestion.
as it would affect the look and tidiness. For example, as illustrated ? We design a novel body shape calculator to determine female
in Figure 2, the sheath dress, whose the style fits very closely to body figure. Empirical results demonstrate that our proposed
the contours of the body from head-to-toe, is best suited to those method significantly outperforms existing methods.
with an hourglass or rectangle body shape, but not good for other ? We construct a benchmark dataset for body shape style rec-
shapes since it will accentuate extra inches and can be unflattering. ommendations. This dataset contains body measurements
With so many rules, limitations and potential style disasters ahead, of 3,150 female celebrities annotated with the correspond-
knowing how to best dress body shape and balance the physical ing types of body shapes and 349,298 images of 270 stylish
characteristics can be tricky business. celebrities annotated with the types of clothing items.
Currently, intelligent fashion analysis has been conducted in- The rest of the paper is organized as follows. First, we discuss
tensively due to the huge profit potential in the fashion industry. related work (Section 2). We then describe our new dataset in detail
Though guide on how to dressing for body shapes is important (Section 3), followed by the overview of our proposed framework
in fashion styling and receiving much attention from fashionistas, (Section 4). Next, we explain our proposed approach in identifying
this issue has been ignored in multimedia science. We, therefore, human body shape (Section 5), modeling the correlation between
propose a novel intelligent fashion analysis framework to model fashion styles and human body shapes (Section 6), and selecting the
the correlation between human body shapes and their most suit- representative styles that best suit body shape (Section 7). Finally,
able clothing styles, which further can be applied to solve fashion we provide experimental results (Section 8) and give conclusions
recommendation tasks. In literature, there are two kinds of fashion and an outlook on further work (Section 9).
recommendation studies. One is recommending outfits that users
may be interested in [33, 44]. The other is recommending fashion 2 RELATED WORK
items that suit to a user-provided fashion item (e.g., boots, cardigan, In this section, we review the related work in terms of (1) Fashion
skirt) [15, 19]. Thus, our proposed work creates a new space in item analysis, (2) Fashion style understanding, and (3) Personalized
multimedia mining and recommendation. style suggestions.
There are three basic research problems studied with this frame- Fashion item analysis. Extensive previous research has been
work - first, a dataset that can reflect the correlations underlying focused on object-based clothing image analysis, such as clothing
body shapes and clothing styles well; second, a learned model that recognition, annotation, segmentation, and retrieval [3]. In recent
can be used to determine the type of body shapes well; and third, a years, a number of models have been introduced to learn more dis-
criminative representation in order to handle cross-scenario varia-
2https://viralyfeeds.com/wedding-dress-fit-body-type-best/ tions [8, 38]. Zhao et al. [45] proposed a novel memory-augmented
Attribute Manipulation Network which to manipulate image rep-
resentation at the attribute level. For clothing annotation, Liu et
al. [27] proposed a clothes dataset with comprehensive annotations
and a new deep model which learns clothing features by jointly pre-
dicting clothing attributes and landmarks. Sun et al. [37] explored a
part-based clothing image annotation approach which takes into ac-
count tag relevance and tag saliency. Both works of Zhao et al. [46]
and Liu et al. [28] addressed the problem of clothing segmentation
and clothing alignment. The result of their works can predict the
positions of functional key points defined on the fashion items. The
works [17, 19, 20] focused on clothing item retrieval research which
can allow a user to upload a daily human photo captured in the
general environment and find similar clothes in online shops. Figure 3: Example of a celebrity profile page.
Fashion style understanding. In addition to some of the tra-
ditional problems, interest in high-level fashion understanding has
been growing in the computer vision community recently. The 3.1 A List of Stylish Celebrity Collection
works [12, 14, 20, 34] explored recognizing and estimating the de- In particular, the celebrities' styles are usually regarded as fashion
gree of fashion styles. These methods allow learning features for references as they hire fashion stylist(s) to help them get dressed
more specific types of images that may be very costly and com- to visually alter their actual body figure. In this work, we therefore
plicated to annotate. Moreover, clothing style understanding can propose to exploit the style of stylish female celebrities to learning
benefit visual analytics of big social data. By modeling the appear- the compatibility of clothing styles and body shapes. We collected a
ance of human clothing and surrounding context, the occupation list of the top stylish female celebrities from a popular crowdsourced
of the person can be predicted in [36]. Chang et al. [2] depict the polling website, i.e., Ranker4, and six popular fashion magazine
street fashion of a city by discovering fashion items that are most websites, i.e., Vogue5, Harper's Bazaar6, Marie Claire7, Glamour8,
iconic for the city. Attempting to directly predict more esoteric mea- and PopSugar9. The top stylish celebrities listed on Ranker are based
surements, such as popularity [11, 41, 42] has also been recently on polling of more than 6,300 voters, while the top stylish celebrities
studied. For style compatibility discovering, Han et al. [9] proposed listed on fashion magazines are selected by fashion editors who have
to jointly learn a visual-semantic embedding and the compatibility expert knowledge in fashion domain. By utilizing these websites as
relationships among fashion items in an end-to-end fashion. a source of information, we can provide an in-depth understanding
Personalized style suggestions. User profiling helps person- of the concept of correlation between body shapes and clothing
alization and has received much attention in the social multimedia styles from the perspective of society as well as fashion experts. We
research fields. Simo-Serral et al. [35] analyzed how fashionable a then recorded all of the names of stylish celebrities listed on these
person looks at a photo whereby advising the user to improve the websites and eliminated the duplicate ones. Finally, we obtained
appeal. Wie et al. [40] intended to explore the inherent relationship 270 names as the top stylish female celebrities for our experiment.
between wearers' personality type and expressive wearing. Sanchez-
Riera et al. [33] proposed a personalized clothing recommendation 3.2 Body Measurement Collection
system through the analysis of user's personal images in his/her We crawled a collection of human body measurements from a
social networks to predict the most likable items. Liu et al. [26] de- celebrity measurements website, http://www.bodymeasurements.org.
veloped an occasion oriented clothing recommendation and pairing We obtained body measurements of 3,150 female celebrities, includ-
system. The system automatically recommends the most suitable ing actress, singers, models, politicians, etc. Each record consists of
clothing by considering the wearing properly and wearing aesthet- the following attributes: name, short bio, type of body shape (i.e.,
ically principles. Though there are several works focus on clothing hourglass, rectangle, round, triangle, or inverted triangle), dress
recommendation system, to the best of our knowledge, our work is size, bust circumstance, waist circumstance, hip circumstance, shoe
the first work devoted to considering body measurement to build size, bra size, cup size, height, weight, and information about breasts
the personal clothing recommendation system. (natural breasts or implants). Note that body shape information
3 DATASET CONSTRUCTION
provided on this website is merely based on visual judgment. Fig-
ure 3 shows the profile of one of the celebrities on the website as
We collected a novel dataset, the Style4BodyShape Dataset3, to an example. We further discarded the data about bio and breasts
enable personalized style suggestion applications that make use of information, as they are not meant to characterize features for
celebrities' style as the knowledge resource. This dataset contains determining the type of body shape.
three types of data: (1) a list of the most stylish female celebrities,
who are known for their sophisticated sense of style, (2) body mea- 4https://www.ranker.com/crowdranked-list/most-stylish-female-celebrities
surements of female celebrities, and (3) stylish celebrities photos. 5https://www.vogue.com/tag/franchise/10-best-dressed
6 https://www . harpersbazaar . com/celebrity/red- carpet- dresses/g8366/female-
In the following subsections, we describe our dataset in detail. fashion-icons/
7 http://www . marieclaire . co . uk/fashion/best- dressed- 2016- 457756
8 http://www . glamourmagazine . co . uk/gallery/best- dressed- women- 2017
3http://bit.ly/Style4BodyShape. 9https://www.popsugar.com/fashion/Most-Stylish-Celebrities-2017-44071762
What is the best fashion style for women? Maje — best for cool-but-sophisticated style Massimo Dutti — best for a muted color palette Stradivarius — best for affordable trends
Title: What dress fits me best? Fashion Recommendation on the Clothing Style for Personal Body Shape
Subject: - Information systems -> Personalization; Recommender systems; Information extraction;
Keywords: Fashion analysis; recommender system; human body shape; clothing style; correlation
Author: Shintami Chusnul Hidayati, Cheng-Chun Hsu , Yu-Ting Chang, Kai-Lung Hua, Jianlong Fu, and Wen-Huang Cheng
Creator: LaTeX with acmart 2018/07/16 v1.54 Typesetting articles for the Association for Computing Machinery and hyperref 2017/03/14 v6.85a Hypertext links for LaTeX
Producer: pdfTeX-1.40.18
CreationDate: Sun Aug 12 12:00:24 2018
ModDate: Sun Aug 12 12:00:24 2018
Tagged: no
Form: none
Pages: 9
Encrypted: no
Page size: 612 x 792 pts (letter) (rotated 0 degrees)
File size: 8134274 bytes
Optimized: no
PDF version: 1.5