Subject
|
Trend in Seattle 2018 - 증강현실을 중심으로 (i.e., “Focused on Augmented Reality)
|
Place
|
Startup Alliance 7F
|
Time
|
March 12, 2018 (Mon) 12:00~13:20
|
Speaker
|
|
Host
|
Startup Alliance
|
Focus
|
augmented reality (or mixed reality)
|
- Lesson
- Lecturer
- personal: Created API at Microsoft for sharing since 2010
- In Silicon Valley, K Group; in Seattle, 창발
- AR
- AR in stories
- Technology in Dragon Ball’s scouter: object recognition (e.g., human recognition from where to where), tracking (the object), texts overlay
- Iron man’s mask: displays missile origin and health system status on screen
- AR research status among big companies
- Google: Tango
- Application example: Kyle Nel’s presentation on Lowe
- Visual positioning system: “Google says VPS makes use of machine learning, computer vision and mapping coordinates…” (source: Mashable)
- Facebook: AR Studio
- Application example: AR Studio
- Technology: Camera effect platform: selfie video app
- C.f., Oculus
- Apple: ARKit
- Application example: IKEA Place
- ARKit: light estimation reflecting real light to objects to make the virtual object real
- Technology
- (horizontal) Surface recognition
- Tracking: smooth
- Light estimation: how much light needs to be reflected
- Google: ARCore
- Technology
- Surface recognition
- Response to real-life event
- Microsoft: Hololens
- Application example
- Education: Case Western University
- 3,000 USD per Hololens device usage compared to xx,xxx USD for dead body isn’t too expensive for med school students
- Industry: Thyssenkrupp’s stairwell
- Technology
- Tracking surrounding environment unlike Google Glass (projecting image just like any computer monitor)
- Hardware: Optical camera 4, many sensors, RGB, HPU (holographic processing unit)
- C.f.,
- Privacy issue: due to 3D point alignment, when uploading an object,
- Safety issue: obstructing wall example
- Other: smart glasses from Intel and Magic Leap
- My Takeaway
- AR application point: use of visual recognition or visual positioning system
- Voice related technology will be needed to augment input on AR devices like Microsoft Hololens
- For object recognition, machine learning technology is needed