Suspected Apple VCSEL supplier reports shipment delay, unlikely to impact 'iPhone 8' availability

Short sellers doing their every best.

AppleInsider said:

VCSEL arrays are a key component in a facial recognition solution Apple is widely rumored to debut in its next-generation handset. Combined with advanced computer algorithms, the system calculates distance to a target using light pulses and time of flight (TOF) measurements, allowing for extremely accurate depth mapping capabilities.

As it applies to iPhone, Apple is expected to leverage depth-sensing tech for user identification and authentication, as well as certain secure features like Apple Pay.

There's an example of what can be done with 3D sensing with Sony's new XPeria phone. They are using just the camera and some photogrammetry techniques to do 3D object capture (1:25):

https://www.youtube.com/watch?v=KDG-pHtNtlI

https://www.theverge.com/2017/8/31/16232720/sony-xperia-xz1-3d-creator-ifa-2017

https://www.youtube.com/watch?v=fqyTTKaNal0

At 5:38 in the following video, they show a good quality scan:

https://www.youtube.com/watch?v=qqPZaHSMrSQ

By using just the standard camera, it takes a while to capture as it has to capture frames and then do image analysis on each frame to determine the surface. A laser scanner can assess the shape much more quickly. Here's a projection scanner scanning an object in 3D:

https://www.youtube.com/watch?v=Qn5e2Zz2EEM

It takes a while to build up and align each separate scan there but the scans themselves happen very quickly and in that example it captures ~500,000 surface points per scan. Putting this in mobile hardware with motion sensors and GPU processing, the alignment and processing of the data can happen very quickly.

This kind of thing is used in movies to create digital doubles of actors:

https://www.youtube.com/watch?v=edG0TF7xpiQ

For the Paul Walker model in Fast & Furious 7, they scanned his brothers and adjusted the models:



http://www.hollywoodreporter.com/behind-screen/how-furious-7-brought-late-845763

With everyone having access to this kind of scanning, it's going to allow for a lot of advanced creative work. If someone is making an indie game or movie, they can do visual effects and scan models to make the process much faster. It has some pretty wild use cases, especially with wearables/glasses. For example, you could scan a family member and even if they died, they could still live with you virtually, you would be able to see them walking around the house. You could have famous people living with you.

It also means AR will have depth masking ability, faster and more accurate tracking and it should allow portrait mode on the front camera as well as face morphing for photobooth-like effects.

The speed of the scan is important for identification. There was a report saying Apple had the scan speed within a few hundred milliseconds:

https://www.bloomberg.com/news/articles/2017-07-03/apple-said-to-test-3-d-face-scanning-to-unlock-next-iphone

It says it captures more data points than touch id. For shape, there may be a way for a 3D printed mask using a face scan to fool it:



but I expect that there will be photographic data too, which would be harder to match with a mask. It'll be interesting to see how they handle the lock-screen. I assume you won't have an option to activate the phone without unlocking to see notifications and the time, it would unlock immediately just by looking at the phone. Maybe they'll have another post-unlock screen for this.

@Marvin - "There's an example of what can be done with 3D sensing with Sony's new XPeria phone...." Awesome comment Marvin! One of the best ever. The use of the technology in movies (Fast & Furious 7) makes it crystal clear what Apple is doing with computational photography.