·If you have a mug, watch, or cell phone on your desk, attach a bounding box only to the watch as seen from Hololens 2.
·The bounding box is projected in real time, and the size changes depending on the size of the clock
·In addition, the bounding box will follow you even if you put the clock on your wrist.
"·I was able to recognize ""desk"" and ""floor"" using the ""space awareness"" function of Hololens, but
"
"I think it is difficult to recognize small things like ""something"" (for example, a clock) on the desk."
·Experimentally, I skip Ray from Head and try to infer the shape of the object using collision determination with the spatial recognition layer.
Space-aware mesh is rough (even high mesh) and can slip through or out of alignment.
If detailed recognition becomes possible, it is assumed that a bounding box can be attached to the target. Do you know any improvement measures or breakthrough measures?Also, I would appreciate it if you could let me know if you have any sample sources for reference.
Device:Hololens2
development environment:Unity
Are you trying to recognize objects from spatial mesh information?
Currently, if it's fine object recognition, I think it's more realistic to make object recognition based on RGB camera image information...
For example, this article contains a sample code: Labeling Toy Aircraft in 3D space using an ONNX model and Windows ML on a HoloLens
618 Uncaught (inpromise) Error on Electron: An object could not be cloned
613 GDB gets version error when attempting to debug with the Presense SDK (IDE)
573 rails db:create error: Could not find mysql2-0.5.4 in any of the sources
916 When building Fast API+Uvicorn environment with PyInstaller, console=False results in an error
© 2024 OneMinuteCode. All rights reserved.