AsianScientist (Apr. 30, 2018) – Scientists in Hong Kong have developed three-dimensional (3D) modeling technology that creates a ‘virtual fitting room’ for users to gauge how well a piece of clothing fits them without having to physically try on the garment.
Since the human body comes in many shapes and sizes, purchasing clothes online can sometimes result in disappointment when the clothing does not fit the buyer well. To overcome this problem, a research group led by Associate Professor Tracy Mok of the Institute of Textiles and Clothing at the Hong Kong Polytechnic University (PolyU) devised an intelligent 3D human modeling technology that digitally reconstructs the shape and size of a person from just two full body photographs.
This innovation allows a customized model to be automatically created within five to ten seconds. The system can accurately reconstruct the 3D shape of a person and extract over 50 size measurements—including bust, waist, hip, thigh, knee, calf and neck, as well as arm length and shoulder slope.
“With an accurate projection of the size and shape of the human body, these customized models will enhance online shopping experience and stimulate growth in online fashion shopping,” said Mok.
The software predicts under-the-clothes body profiles of subjects based on over 10,000 input images where the body profiles are covered in arbitrary clothing, including tight-fitting, normal-fitting and loose-fitting clothing.
Then, deep learning technology was applied to segment the human body image from the background, improving the robustness, efficiency and accuracy of the shape modelling. Finally, the researchers created a mobile application using the automatic shape customization technology, which can be downloaded from the App Store or Google Play.
The discrepancies in tight-fitting and loose-fitting clothing were found to be less than 1 cm and 2 cm respectively, which are precise enough to fulfil the specific requirements of the clothing industry for fashion applications and comparable to body scans.
“The output models can also enable customers to visualize try-on effects before making purchases in online stores. This frees us from the limitations imposed by taking body measurements physically, helping customers to select the right size in online clothing purchases,” added Dr. Zhu Shuaiyin of PolyU, who is a co-author of the study.
———
Source: Hong Kong Polytechnic University.
Disclaimer: This article does not necessarily reflect the views of AsianScientist or its staff.
