ObjectDetector
December 2025
ObjectDetector was a 3rd-year university project to begin an interest in prototyping interfaces for XR which also utilised ML vision models - in my case, this example utilises Meta's Quest 3 / Sentis YOLOv9 sample, reverse engineered to be a HUD with a per-class selection to allow a user to select from items that are "detected" from their surrounding environment, to be able to highlight where they are.
The input mechanism also utilises the thumb microgestures seen in the Meta SDK + Ray-Ban Display glasses' Neural Band for fast selection and navigation, with an option to also focus vision on only selections without the HUD UI.
Whilst only a short 2-month long project, it paved the way for my later works such as Paper Plane Folder and beyond…