F. Donadoni et al., "PS for You: A Contactless, Remote, Cloud-Based Body Measurement Technology Powered by Artificial Intelligence", Proc. of 3DBODY.TECH 2021 - 12th Int. Conf. and Exh. on 3D Body Scanning and Processing Technologies, Lugano, Switzerland, 19-20 Oct. 2021, #31, https://doi.org/10.15221/21.31.

Title:

"PS for You": A Contactless, Remote, Cloud-Based Body Measurement Technology Powered by Artificial Intelligence

Authors:

Francesca DONADONI 1, Jayne MECHAN 2, Colette JOHNSON 1

1 The PS Collective, New York City, USA;
2 Manchester Metropolitan University, Manchester, UK

Abstract:

Contactless and remote measurement technologies have recently been at the core of the digital shopping experience, not only for their potential to provide better customization but to revolutionize the way customers shop - from minimizing returns due to misfits to reducing environmental impact by producing on-demand. For small-batch retailers, this is key in ensuring minimal waste and streamlining the consumer journey. From the consumer's point of view, fit is also an unaddressed issue. After conducting user interviews on 60+ women, we identified 70% of respondents reported difficulties finding proper fitting apparel by designers. Reports by Mintel [1] and Deloitte [2] also confirm that the issue of fit is paramount in the industry. Here we introduce "PS for you", our AI-powered solution to provide a fast and scalable platform for designers to obtain reliable measurements and for shoppers to find the right fit.
Our solution uses computer vision models paired with a proprietary layer of post-processing algorithms to extract personalized measurements from photos that the user can upload from any device. The solution is deployed on a serverless CPU using Amazon Web Services (AWS) for easier interoperability and faster run times.
To test the effectiveness of this approach, we collected 40 images from 20 volunteers of different body shapes and sizes (including petite, short-stature, and athletic). The system was able to detect measurements for waist, hips, and bust with an average difference of less than 3% with respect to tape-measured data. The median difference remained <1in, a significant improvement which is 3x more reflective of true size than the current universal sizing methods [3].
Working on this technology allows us to foster a new connection between shoppers interested in customization, sustainability, and comfort and designers who are interested in decreasing returns, eliminating waste, and understanding more about their consumers. Technology will not just help but can create opportunities to build new systems, without heavy manual support that connects and fosters experiences anywhere in the World. By using the right technology and data, responsibly, we're building a product that will open up opportunities in connected markets as well as enhancing existing ones.

Keywords:

Machine learning and artificial intelligence for 3D body modeling and processing, Portable human body measurement systems, Apparel and fashion, Custom apparel, Digital anthropometry

Details:

Full paper: 2131donadoni.pdf
Proceedings: 3DBODY.TECH 2021, 19-20 Oct. 2021, Lugano, Switzerland
Paper id#: 31
DOI: 10.15221/21.31
Presentation video: 3DBodyTech2021_31_Donadoni.mp4

Copyright notice

© Hometrica Consulting - Dr. Nicola D'Apuzzo, Switzerland, hometrica.ch.
Reproduction of the proceedings or any parts thereof (excluding short quotations for the use in the preparation of reviews and technical and scientific papers) may be made only after obtaining the specific approval of the publisher. The papers appearing in the proceedings reflect the author's opinions. Their inclusion in these publications does not necessary constitute endorsement by the editor or by the publisher. Authors retain all rights to individual papers.

Note: click the + on the top left of the page to open/close the menu.