Body Shape Recognition for Online Fashion

Lead Research Organisation: University of Surrey
Department Name: Vision Speech and Signal Proc CVSSP

Abstract

Abstracts are not currently available in GtR for all funded research. This is normally because the abstract was not required at the time of proposal submission, but may be because it included sensitive information such as personal details.

Publications

10 25 50
 
Description The primary advance achieved in this project is the introduction of a method to measure and classify human body shape from a single image capture on a mobile device.

This technology has been deployed and evaluated as a mobile app to allow clothing recommendation for online retail.



Key contributions of the research include:

(1) Novel learnt representations of human body shape and pose variation from a database of 3D body scans.

(2) Robust reconstruction of human shape from a single image with unknown camera calibration.

(3) Statistical model based human segmentation from images of people in natural scenes.

(4) Classification of human body shape using image-based reconstruction.

(5) Measurement of human shape dimensions from a single image for clothing size recommendation.

(6) Development of a mobile app allowing body shape classification and measurement from a single image captured on a mobile device.

(7) A low-cost system for human body scanning using multiple Kinect RGBD sensors to capture a database of 3D body shape.



The application for human body shape classification and measurement has been integrated with an online system for clothing recommendation developed by the London College of Fashion as part of the project.

This system has been evaluated for in-store and online clothing recommendation with high-street retailers.
Exploitation Route Primary application: Online clothing retail.



Secondary applications: Body shape measurement for health and fitness applications.

Technology for estimation and classification of human body shape from a single image has attract considerable interest for commercialisation from online and high-street retailers internationally. The mobile app is currently under discussion for licensing. This research is also the focus of further development and technology transfer.
Sectors Digital/Communication/Information Technologies (including Software)

Retail

Security and Diplomacy

URL http://cvssp.org/Personal/AdrianHilton
 
Description This project introduced methods for body shape measurement from a single image. This method was used to develop a mobile app for clothing recommendation for use in online retail.
First Year Of Impact 2013
Sector Creative Economy,Digital/Communication/Information Technologies (including Software),Retail
Impact Types Cultural

Societal

Economic

 
Title Multiview and 4D Video 
Description Multiple view video datasets and reconstructed 4D models 
Type Of Material Database/Collection of data 
Provided To Others? Yes  
Impact Data available is used by over 300 research groups worldwide. It is considered the benchmark dataset for research in this field. 
URL http://cvssp.org/data/cvssp3d/
 
Description BodyMetrics 
Organisation Body Metrics
Country United Kingdom 
Sector Private 
PI Contribution Methods for body shape measurement from images
Collaborator Contribution Access to data of body scans
Impact research publications and tools
Start Year 2014
 
Description Numerion 
Organisation Numerion Software Limited
Country United Kingdom 
Sector Private 
PI Contribution Validation of physics based cloth simulation against real cloth measurements
Collaborator Contribution Physics based cloth simulation expertise
Impact Physically validated cloth simulation tools
Start Year 2014
 
Description Sony Broadcast and Professional Europe 
Organisation SONY
Department Sony Broadcast and Professional Europe
Country United Kingdom 
Sector Private 
Start Year 2004