Case Studies in Scanning Humans to Create Customized Products Thomas Tong, 3D3 Solutions
Jan 19, 2015
Case Studies in Scanning Humans to Create
Customized ProductsThomas Tong, 3D3 Solutions
Overview
• 3D3 Solutions has created 3D scanners that scan humans to create customized products
• Lessons learned for making 3D scanning usable by the general public
• 3 different cases - Feet, Faces and Joints
Goal
• Best practices for 3D scanning technology to solve human measurement problems
• Understand value proposition of moving 3d scanning from back-end systems
• Discuss risks and pitfalls
• Tackle production issues
Key Issues 1
• Scanning humans has different requirements than mechanical parts
• Movement• 3D scanning working
outside of controlled settings
• Used by untrained and non-technical operators
• Scanning speed and robustness
Key Issues 2
• The cost factor is challenging– Competition is traditional
tools– better, faster, or cheaper– Digital solutions has
different trade offs versus traditional tools
• Overcoming social fears
Methodology
• Custom build 3D scanners for each particular application
• Scanners prototyped and tested with end-users every step of the way
• Adapt available technologies, mix and match old and new techniques
• Modify the parameters of the problem
Case Study 1 – Feet
• Semi-custom to fully custom orthotics
• Scanning feet– Plaster cast and foam replacement
• Simple value proposition. Plaster and foam requires more time and physical handling costs– Digital scanning is simple, clean and fast
Screen Shot
Screen Shot
Case Study 1 – Lessons Learned
• Scans asked to be highly reduced– High data sets = more storage, processing, and
transmission time• Less is more. The right measurements, not
more measurements• Scans data are used to look up into a model
library, additional data is not needed• Accuracy requirements are low. Ease of use
requirements are high• Time is money. Faster is better
Case Study 2 : Faces
• Faces– Medical
• Cosmetic surgery• Burn
– Visual effects• Digital Avatars• 3D Models
Screen Shot
Case Study 2: Lessons Learned
• There will be holes in a single scan – Solutions:
• Automatic - hole filling• Use replacement parts (No ears? No problem)• More cameras or scanners
• Glasses can be an issue• Hair in general does not scan that well
– Even if data is captured, a polygon mesh represents hair poorly
– In visual effects or games you would replace with a hair simulation
– For manufacturing, mesh may be okay
Case Study 2: Lessons Learned
• White light and laser safety– Rated to be safe– People are very skeptical, especially athletes
• Faces are the hardest thing to keep “still”• Different industries
– Visual effects, artists want to do lots of editing, can deal with lots of polygons
– For manufacturing, take the face or head scan and blend it in to a pre-existing model for less cleanup
Case Study 3 – Joints
• Bracing Products– Knees– Elbows– Ankles
• Data acquisition from multiple directions
Case Study 3 – Lessons Learned
• A fast 360 may requires multiple cameras– Systems that sweep around the joint will work– While humans are not rigid, general shape
remains unchanged
• Automated alignment can be challenging
Screen Shot
End User Issues• People don’t know what to
expect• There is significant reluctance
and fear– People worry about being
digitally “copied” especially women
– Social fears to the 3D scanning like early photographers
• Everyone moves all the time– Faster, higher cost, higher
quality scanners– Support the persons (Wall,
chairs)– Live with the noise
Lessons Learned
• Plan your scanner output to match required manufacturing input
• Fully automated solutions are difficult, but worth the effort
• Robustness is critical • 100% coverage near impossible• New methods will rarely be “cheaper” than
old methods. – The ability to digitally measure, analyze and
deliver data is valuable
Easy Wins
• 3D scanning is “cool”’
• Value proposition is straight forward– Existing solution is error prone, slow, difficult,
and requires shipping– Digital solution is fast, accurate and can be
digitally delivered
• Accuracy requirements are orders of magnitude lower than manufacturing
Future Steps
• Making simple and robust “1-click” scanning systems
• Look for end to end value, not just 3D scanning
• Develop digital measurement, analysis and information delivery platforms.
• Get more correct measurements more easily at a lower cost