STUFF
Body Analysis: Privacy
Company: Advanced Health Intelligence (AHI)
Complexity: 8.876/10
Fun factor: 9.976/10
What did I do?
- Discovery, research, designs, iOS prototype (Xcode), presentation, documentation.
What is it?
- A psychological approach to address privacy concerns, by not showing the video, but rather a real-time silhouette of their body.
- Bonus features: A competitor differentiator through strong visual appearance and customer branding options.
Masking humans: It looks like those apple iPod ads, doesn’t it?
The problems
The scan requires people to be in form fitting clothing. We found that some people just got into their underwear, presenting some problems when viewing themselves on a screen:
- They question the privacy and security.
- They don’t like what they see (themselves, partially naked) and find that invasive and then don’t like doing it (basically a mirror).
How is it done?
Real-time Matte Effects using Apple VisionFramework
Mask modes: the higher the accuracy, the slower the performance.
Major challenges
- Device performance impacts the accuracy of the masking layer.
- On low-end devices, this may also impact the on-device inspection, slowing down the overall experience.
Check out these clean joints…
Major wins
- Full image/video backgrounds
- Clean real-time joints by using coloured masks, the joints can be enlarged, create a high contrast, and improve legibility from a distance.
- Reduced visual clutter in determining limbs into outline.
- Alternate capture through fit-frame modelling (no outlines).