Your Cart is Empty
Delivered every working day, our email newsletter inspires the world's greatest sports innovators to create the future of sport. Enter your email below to subscribe - it's FREE!!!!
I know I'm not saying anything new when I say that we come in all different shapes and sizes. And athletes in particular are no different in that respect, even within the same sport there can be a huge range.
Just think of the difference between a prop and a winger in rugby!
Add to this the issue that manufacturers ideas of what is supposedly the same clothes size can vary enormously, so buying clothes that fit correctly (an essential for sports!) can be a minefield.
But what if you could scan your body, completely accurately with your phone?
Game changer right?
You would know that that essential piece of kit was going to fit like it was especially made for you...
Rosa Porcar, Director of Internationalisation and Innovation at the Institute of Biomechanics in Valencia, Spain, has been working on this very thing and kindly gave us permission to republish her article here so that Kit Radar readers can find out more.
The usefulness of having personal and accurate 3D body data is undeniable for any process involving product fitting or assessing body shape evolution.
Examples of use of 3D body data for product fitting or made to measure are: Apparel industry, Personnel Protective Equipment, Sport equipment, Orthopaedic braces and orthotics.
Assessing body shape evolution is a must when quantifying the impact of a diet, a medical therapy, a workout programme, …
Having said that, how do we obtain accurate individual 3D body data?
Body scanners can acquire the body surface of minimally dressed human bodies and measure them digitally. Their technology has evolved and their cost has substantially dropped. They are more portable and precise. However, there is still a need for a specific physical device, and bulky systems are still required if measurement precision is needed.
Thanks to large databases created using 3D body scanners, IBV has developed a lower-cost technology based on data-driven 3D reconstructions (D3DR) to estimate human body models from 2D images (2D3D). D3DR solutions have much lower hardware requirements than current 3D scanners. This solution is suitable for running in mobile devices such a smartphones or tablets, making possible to have a 3D body scanner in your smartphone.
Therefore, some questions arose:
Let’s see the answers.
IBV has developed an Android App (2D3D), where user is guided to take two photographs:front and side. Instructions are provided during the process. It can be done anytime and almost anywhere in less than one minute. Very few conditions are required.
Height and weight, gender and birth year are asked.
Then, two photos (front and side) must be taken. Two silhouettes appeared to guide you to find the right distance.
Using AI, the output is a 3D Body Avatar. Clean, closed, a mesh with 50K points. Single-surface watertight 3D model. A high-resolution mesh ready to interface other CAD and simulation applications.
The mesh is provided together with a list of up to 100 measurements defined according to reference standards (ISO 7250, ISO 8559: Body volume, Cervical height, Crotch height, Mid neck girth, Shoulder width, Shoulder length, Shoulder breadth, Bust/chest girth, Underbust girth, Waist girth, Hip girth, Arm length, Upper arm girth, Wrist girth, Max thigh girth, …)
Living bodies 3D scanning is complex. While 3D scans are static and rigid, living bodies are soft and articulated. When scanning a person several times, the pose and muscle contraction, among other factors, may change slightly, giving different scans and measurements for the same person… but all are right!
Reliability can be described as the likeliness of someone repeating the experiment and getting the same result. In order to calculate the reliability of our 2D3D system we did it in comparison to other mainstream body measurement systems.
For that, we measured 77 subjects (males and females covering a wide range of body height and weight) with the 2D3D app running on a Motorola Nexus 6 and with Vitus XXL body scanner from Human Solutions GmbH (Laser tech). Each subject was scanned and photographed twice with repositioning, getting in and out of the scanning/photographing space at each repetition.
To assess reliability we estimated, among others, Mean Absolute Difference (MAD). MAD is the average distance between each data value and the mean. It is a way to describe variation in a data set.
We compared the obtained MADs 2D3D vs Laser tech vs measurements taken by experts with traditional anthropometric methods (TA)
2D3D provides MAD very close to Laser tech and within the range of best results reported by studies of TA.
See table below.
As a result, we have a very reliable system, but reliability is not enough by itself (an experiment which you can repeat many times and get results close to others is very reliable but it could be the same wrong result each time!)
Therefore, we need to compare the obtained values with the true values. We are going to compare values from 2D3D to values from Vitus XXL body scanner from Human Solutions GmbH (3D3D).
Mean Absolute Error (MAE) in centimeters between actual scanner and 2D3D is presented in the table below and compared to Maximum allowable error established in ISO 20685:2010 - 3-D scanning methodologies for internationally compatible anthropometric databases. 2010.
MAE of 2D3D with respect to Laser Tech is, for most measurements, within the threshold established by ISO 20685:2010 or very close to it.
Our 2D3D reconstruction from two images can be implemented in a regular smartphone thus making it accessible to millions of people without the need of specific hardware or skills.
2D3D provides visually accurate 3D models and the reliability of measurements obtained from these models is comparable to body scanners.
This article is part of our Kit Radar science series, if you have some interesting sports technology science please get in touch (info@kitradar.com).
Like the content and want to be ahead of the curve? Enter your email below to subscribe - it's FREE!!!!
