SnapML Proposal 2020
Augmenting how we navigate
time
through machine learning.
TethARball!
“TethARball” transforms users’ surroundings into vibrant interactive AR tetherball experience. This machine learning-powered AR game enables players to kick around a tetherball with their own feet. Machine learning models are used to map concrete-like terrain to the ground, detect the collisions between feet and the virtual tetherball, and even play together with someone else by detecting their feet too!
“TethARball” is in AR, uses ML, and can be optionally played with 2 players as a multiplayer experience. Balls could be customized from the default tetherball, such as a blazing fireball.
Machine Learning Models
A futuristic fusion of Style Transfer, Ground Segmentation, Foot Tracking, and Object Detection
Style Transfer
Ground Segmentation
Foot Tracking
Object Detection
AR | ML
By combining Augmented Reality and Machine Learning, the interactive games of the future—-games that are personalized to users’ own environments—-are now possible. ML is the slice of data that will manifest vibrant customized AR worlds.
TethARball democratizes the need to be connected to reality. For everyone.
