You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
There is huge room of performance improvement because libdepthengine.so uses inefficient OpenGL shaders to resolve depth images. The libfreenect2 project https://github.com/OpenKinect/libfreenect2 had reverse engineered the OpenGL shaders (and even the USB wire protocol and formats(!), which was reasonably open sourced later) used by Kinect v2, and I rewrote and optimized the shaders in CUDA, showing ~100% improvement of frame decoding time https://github.com/OpenKinect/libfreenect2/wiki/Performance, which means the OpenGL shaders are inefficient and it is very useful to use CUDA to decode the depth images for resource constrained embedded platforms.
I think there is no secret sauce in the OpenGL shaders.
Do you have a corporate contact? In case you feel this issue is not convenient for public disclosure.
The text was updated successfully, but these errors were encountered:
https://github.com/orbbec/OrbbecSDK_v2/tree/main/extensions
Can I ask why this is not open source?
I understand Microsoft abandoned Azure Kinect and licensed the IP to Orbbec. But is there a specific reason why this part was not open sourced?
There are two benefits of open sourcing this:
I think there is no secret sauce in the OpenGL shaders.
Do you have a corporate contact? In case you feel this issue is not convenient for public disclosure.
The text was updated successfully, but these errors were encountered: