This document discusses how computer vision will change augmented reality and the world. It summarizes how science fiction concepts like space travel, weather control and television were envisioned before becoming reality. Augmented reality aims to abolish the interface between digital and real worlds by turning everything seen into an augmentable medium. The document outlines how sensors define reality and how computer vision is crucial for augmented reality. It provides examples like a Yelp app that overlays ratings and names on storefronts and discusses how sensor data can be imperfect. It explores how computer vision can replace hardware by recognizing text, tracking motion, reconstructing 3D scenes and more. The conclusion is that cameras will define future augmented reality and turn more science fiction concepts into science fact
8. Yelp Monocle
ā¢ Built by yours truly
ā¢ Point your phone down
the street, see ratings
and names overlaid on
store fronts
ā¢ Reality deļ¬ned by
compass, accelerometer
and GPS.
9. Sensor Data is Dirty
GPS coordinates are
accurate to ~+/-50m (i.e.
an entire city block)
This is where I actually am
18. Even Simpler Things
Need Computer Vision
Square
Valued at $40 million before launch
Provides a audio jack dongle to read cards
Struggling to meet hardware demand
19. Why bother with hardware, when 45
lines of code and a camera will do
the same?
Credit Card Recognition circa Last Night at 2 a.m.
20. And Many More
Logo and Storefront Recognition
Barcodes/QR Codes
Medical Image Diagnostics
(Cool new Yelp things)
21. The Revolution Is
Happening Now
Real-time camera access in iPhone 4.0
(public release in June)
HD Video decoding chips can be co-opted for
computer vision
22. In Conclusion
Cameras will deļ¬ne our future augmented reality
In doing so they will further the software crusade
The restraining factor is not technology, itās is the
ambition to execute.
Letās make Sci-Fi become Sci-Fact
Thanks for listening! (follow @newhouseb for more)
Editor's Notes
Hi
more than just putting a dinosaur by the golden gate bridge. More than using your orientation as search parameters.
But this was yesterday, tomorrow, it will be all about the cameras (change the polaroid photo)
Need a better picture (NIST examples).
Replacing? Humans reading, dictionaries, pen readers