WebXRMeshDetector should support specifying a level of detail for reported meshes
The UWP OpenXR runtime just released a preview 3 extension around scene understanding that introduced some changes to how mesh detection works.
- Scene objects can now have parent objects. We may want to update how detected planes and meshes are reported so that the scene object parent id can also be determined.
- Mesh detection now has separate computation calls for collision meshes (lower poly) vs visual meshes (higher poly). We will want to update XRGeometryDetectorOptions to allow for specifying a preferred level of detail: https://github.com/BabylonJS/Babylon.js/blob/4d5ae77dd6d617bd99fae0c29098708a8078e619/src/LibDeclarations/webxr.nativeextensions.d.ts#L55
It likely isn't worth starting on this work until the scene understanding openxr extension is out of preview, which will likely happen over the next couple of months. Until then the API design may change. However, if someone starts looking into implementing ARKit/ARCore support for mesh detection prior to the permanent extension being available, we may want to rework the meshdetector/planedetector api surfaces to support these changes.
Any update here?
Any update here?
Hey Raanan, I no longer work at Microsoft. You may want to reach out internally to folks now maintaining webxr HoloLens 2 support.
@bghgary - can this be re-assigned?
Closing as no activity. Feel free to reopen