ARKit vs ARCore in the Task of Baggage Measurement
summary
Compares ARKit and ARCore by actually building a baggage-measurement feature: floor plane + accumulated feature points form a point cloud, filter out points within ~5–10cm of the floor, compute convex hull (QuickHull) and minimum bounding rectangle (rotating calipers), take max height for volume, render a live box (iOS uses ARKit scene tools, Android falls back to OpenGL ES2). Finds ARKit better at short range, ARCore better at long range, typically uses ~100 feature points, and reports ~1.5 cm average error; fails on shiny floors, glass surfaces and close furniture.
AI-written summary · about
chapters (8)
- 0:00Introduction and context
- 1:12The baggage problem
- 3:58First attempt at a solution
- 9:37Second iteration and floor
- 14:06Point cloud construction
- 18:09Calculating object volume
- 22:02Rendering and results
- 26:57Industry context and Q&A
related talks
58mImplementing Augmented Reality (AR) in Apple's ARKit
24mARKit Introduction
35mAugmenting Faces and Images
15mDepthLab: Real-time 3D Interaction with Depth Maps for Mobile Augmented Reality
42mARrrrg The Google AR Android Demo Application deciphered
15mDepthLab: Real-time 3D Interaction with Depth Maps for Mobile Augmented Reality
39mNow smile (also in 3D)! Exploring AR, ML and Camera related APIs on Android
06mUnravelling Spatial Privacy Risks of Mobile Mixed Reality Data