Clothing Comfort Predictability via 3D Body Scans
Keywords
No Thumbnail Available
Authors
Issue Date
2018-07-01
Language
nl
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Prediction of clothing comfort is a challenging problem
in virtual fitting. Most of the solutions tried so far are based
on external devices or sensors. We propose a novel method to
predict clothing comfort based on only 3D body scans.We created
a small 3D scan database of 12 participants’ lower body dressed
with jeans on leggings and just with the leggings. This dataset
was used as input for several Neural Networks and machine
learning algorithms. The input was the area of cross sections
taken perpendicular to y-axis from particular regions of the
scans (knee, thigh, crotch, hip). Additionally, we created an image
dataset to be used as input for a Convolutional Neural Network.
We used Neural Networks or the Machine Learning Algorithms
to predict clothing comfort evaluated by the participants after the
trials of the jeans. Results support that clothing comfort problem
can be handled as a classification or an image classification
problem and lead us to the possible adjustments to be able to
predict clothing comfort with 3D body scans precisely.
Index Terms—Clothing Comfort, Virtual Fitting, 3D Body
Scanning, Image Classification
Description
Citation
Supervisor
Faculty
Faculteit der Sociale Wetenschappen