SPOTLIGHT ON:
CHIEF DATA SCIENTIST, DR. JIE PEI
What's Wrong with Traditional Bra Fitting?
We sat down with one of the most brilliant minds at FIT:MATCH, Dr. Jie Pei,
whose inventions have shaped the DNA of our company’s solution. Our Chief
Data Scientist walks us through her extensive research over time in breast
shape (including a Masters and PhD at Cornell University plus 3 patented
methods). She also shares how this 3D problem demands a 3D solution.
Clothing fit is a big challenge for product designers and brings frustrations to consumers. The insufficiency of traditional
body measurements (circumferences, widths, lengths, etc.) in describing body shape has been continuously discussed
by researchers, yet nothing fundamental in apparel sizing has changed since the 1940s. Amazed and fascinated by the
variation in female body types, I dedicate my career to solving fit problems by:
1) studying body shape, especially breast shape, in 3D and exploring methods to evaluate shape other than the traditional
measurements, ultimately to optimize sizing
2) exploring methods to revolutionize standard sizing methods
Among all body segments, the Moreover, most upper body fit issues (for items ranging from blouses to body
breasts have the most armor) occur in the bust area. Therefore, I primarily focused on breast shape
complicated 3D shape with during my seven years at Cornell. 3D scanning has been commonly used to study
tremendous variations among body shape for product development purposes. My skills in programming allow me
individuals. to experiment with new ways to view and analyze 3D body scans.
Fig 1. How my research has evolved over time
For my master thesis, I proposed categorizing breast shapes based on ratios and angles, using a vast total of 66
nontraditional measurements. In my PhD study, I proposed the use of 2D topographical plots to evaluate breast
shape and breast asymmetry. In the final years of my PhD, I proposed the 3D mapping method that takes advantage
of all the points on the scan surface to capture the shape information in 3D. I was also able to integrate the 3D
information to the design of a sizing system and I proposed solutions to select fit-models and to recommend sizes to
consumers. Thus, my work has progressed from 1D (linear) measures to full 3D analysis (Fig.1).
3D contains far more information than 1D or 2D. 2D information can be easily extracted from 3D. But the other way
around is much harder, because there could be thousands of different shapes for a certain combination of
measurements. Similarly, it’s hard to distinguish sizes solely by measurements. 36C and 36D, for example, will have
large overlapping, if not completely the same, measurements. Breasts have such complicated shapes that 2D is not
enough (study after study has proved this), let alone to rely on 2D to account for breast density and many other
factors. 3D, on the other hand, accounts for all the factors.
In short, a 3D My 3D method is the first true 3D system ever developed to categorize
question requires body and breast shape. It includes a new way to align body scans for
mapping, which allows for a standardized reference point and facilitates
a 3D solution. the automation of handling 3D body scan data. It moves away from the
total reliance on measurements, and allows customers to find their sizes
accurately based on their shapes. (Details of the 3D method have been
fully documented in our recently granted US patent).
Fig 2. Same bust circumference, very different shape
As a result of failing to incorporate breast shape information, many bra sizing systems are fundamentally flawed.
Brands struggle to make their sizing system more inclusive to the ever more diverse body shapes, satisfying the
needs of the smaller and larger chested customers. The existence of “sister sizes”, where a woman who wears
34C can potentially also fit into a 36B or even a 32D bra, brings confusion to customers.
Up to 85% of women wear the wrong size bra.
In addition, there is a lack of standardization in bra sizing, leaving bra sizes/fit to vary greatly
by designs and companies. Each company has their own rules of converting the bust-
underbust difference into letter graded cup sizes, and the offset can range from 0 to 6 inches.
Lastly, the practice of Vanity Sizing, where companies deliberately alter size labels to flatter
smaller-busted consumers by indicating that bigger than actual sizes are the right fit,
exacerbate the inconsistency. Despite all this, and the increased diversity in breast shapes,
bra sizing systems have not changed fundamentally since 1935. No major development of a
scientific and practical method that improves bra sizing fundamentally has been developed. It
is a hard problem to tackle, yet mobile body scanning technology and the emphasis on
shapes can be a game changer.
Being a girl who loves challenges and aims to fix real-life issues, I developed the 3D
mapping method specifically with all these pain points in mind. And the patent itself is not
just a better size recommendation solution. For example, finding ideal human fit models
has always been a challenge for the industry. There was no reliable selection criteria for
fit models. My 3D solution, however, takes advantage of the company’s existing
customer base (or their targeting demographic), builds out a shape matrix by collecting
scans, and for the first time, offers a shape-based solution to help the company select fit
models from this customer base. Moreover, with the power of LiDAR mobile scanning
technology and the ability to collect tens of thousands scans/shapes every minute
through smart phones, we can inform bra designers, pattern makers, manufacturers on
critical shape-based fit information.
Fig 3. Patented Shape Matching Algorithm
At FIT:MATCH, we can help fix the
traditional bra sizing systems, and bring
revolutionary change to the industry. This
is my ultimate goal and passion.
FIT:MATCH aligns with my vision, and has
all the right resources to make it happen.
My work has grown immensely since I
joined the team a little over a year ago.
Our team inspires me and sparks
innovation. This has encouraged me to
invest the best outcomes of my 7 years of
hard work at Cornell into FIT:MATCH.
The best is yet to come!