In data scans, there seemed to be dramatic differences

In the case of the TC2-19 3D body scanner, the two test subjects performed only one full body scan each. This allowed us to captured additional data points, which were used for comparison with the Size Stream data and to further assist in the development of usable avatars.

 

When comparing the various measurements from the Size Stream data scans, there seemed to be dramatic differences from one data set to the next. Even from the left sides to the right sides of the test subjects, the measurements were found to be dramatically different at certain data points.

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!


order now

These anomalies raised the question of the reliability and acceptability of the scans.

 

We then investigated the difference between the Size Stream scan body measurements and the target subjects using a traditional tape measuring approach, which too yielded widely varying differences. Findings indicated that significant differences existed between circumference measurements calculated by the body scanner and those gathered from a tape measure approach. Some of the differences could be attributed to where the measurements were being taken, but even in those cases some of the measurements were dramatically different. In addition, the errors that we discovered in our scans were corroborated to have been seen by other faculty members from the production management department who also used the same body scanners.

 

In our research we were not only looking for the scans to deliver accurate measurements, but also produce highly detailed 3D models that could be used as avatars of the test subject’s bodies. In actuality, scanners produce clay like, organic silhouettes. This is because body scanners decimate points and therefore, are not made to accurately capture the body. Our research revealed that the images produced from 3D scanners are not usable to directly start draping on. This led us to seek out advice from the experts at Alvanon. Our research team met with Susana Charm, Marketing Manager and Daniel Naranjo, Creative Director at Alvanon to discuss their virtual avatars and the process by which they are produced. Alvanon is the leading global apparel business and product development consultancy firm in the industry. They provide fully integrated solutions to their clients that improve many of the industry challenges apparel businesses face today. The company has a global team of apparel experts who equip the top fashion retailers, brands, designers and manufacturers with growth, customer engagement, product development and supply chain strategies.

 

Alvanon is currently testing all of the mainstream body scanner technology in the market today. The data will be used for a size study they are currently conducting in Hong Kong. The company is currently working with the TC2 scanners. They have also worked with the Human Solutions laser scanner and have worked with other scanners in the past.

 

Alvanon, who produces 3D avatars that replicate digitally the dress forms they develop and manufacture, apprised our researchers that 3D scanners are designed only to measure quickly.  “3D scans represent fit models not a form for designing, the scan has to be further cleaned up,” according to Daniel Naranjo. “Most 3D scanners are designed only to capture the body measurements accurately, these scans cannot be used as a form for designing.” He also said “The scans require extensive post processing to turn it into a usable tool. No matter what is done to scan the body, there will still have to be manual cleaning to remove anomalies.” This is because scanned bodies usually have flesh touching flesh, such as in the armpit area and in between the legs. In order to turn it into a usable tool the arms would have to be milled out in such a manner that the excess is removed yet the circumference is still maintained, that way the arm is away from the body but still holds its original measurements.

 

Scanning Technologies

 

Different technologies are used to build 3D-scanning devices; each technology comes with its own limitations, advantages and costs. In our case, both the Size Stream SS20 and the TC2-19 scanner technology use Infrared Depth Sensors, which is an enhanced structured light scanning technology.  But, 3D body scanners that use lasers are the ones that might work best for creating an actual figure. 3dMD scanners develop anatomically-precise 3D and “temporal-3D” (4D) images. 3dMD is capable of scanning live subjects to the degree of image accuracy required for planning in the areas of full-body anthropometrics and size/fit. These surface imaging systems use the kind of sophisticated software required to support serious applications in healthcare, biometrics, ergonomics, human factors, research, and size and fit. The migration from laser scanners to 3dMD body Systems will help optimize the ergonomic fit of equipment and design creations.

 

3dMD systems ensure an optimum Episode of Encounter with every subject by combining ultra-fast scan speeds (1.5ms) to ensure anatomical-3D precision with every capture and workflow efficiencies to handle the challenges associated with high-throughput environments. This ability enables 3dMD images to support 3D printing applications and new concepts in sporting and leisure apparel with minimal labor costs, no measurement error, and a positive consumer experience. (Citation) http://www.3dmd.com/applications/size-and-fit/

 

Importing the MEP data

 

When the body scans could not be successfully used as avatars for draping purposes, the decision was made to apply the measurements defined within the MEP (Measurement Extraction Parameters) from the body scan to modify the parametric avatars already in existence in the Browzwear solution. From there we would create avatars that mirrored our two test subjects. This led to our second revelation about working in the 3D space. Many issues occurred while attempting to import the measurements from the scanned data onto the avatars used in Browzwear.

 

When attempting to apply the data to the parametric forms we found that there was no reliable way to extract and align the waist, hip, bust, inseam and other body measurements from the scanned data to the avatars in the system. Editing of the 3D human avatar is controlled by a few semantically meaningful parameters: 12 biometric attributes control the body shape, and 17 bones control the pose. Browzwear’s parametric editing system is based on a revised SCAPE model. However, because the parameter table is locked to a 3D body-aware warping scheme to reduce the possibility of producing an implausible and unnatural looking figure, it also prevented our researchers from creating a body shape which was similar in nature to the test subject.  It was later discovered that there was a huge discrepancy between where the POMs (Points Of Measurement) were occurring in the scan data and where the measurements were being placed on the avatars. This made it particularly difficult to duplicate the actual silhouette of the scanned subjects. Being able to precisely input the measurements from the scanned data or from a fit model into the system is key to the development of an accurate and functional computer-based avatar to work on.