Open pablovela5620 opened 1 year ago
Good idea! Let me update the dataset
I cannot seem to be able to optimize the original poses produced from ARKit. The poses I have are obtained from camera.transform. However, they are inaccurate whenever I try to visualize them. I would highly appreciate if you could help me with this.
Hello @likojack, honestly great work!
Can you share with us please the ARKit poses! I want to test it on my vSLAM system as well!
Thank you.
I'm assuming the poses included in the dataset are those obtained after bundle adjustment optimization. Would it be possible to include the original poses produced by ARKit? This would help with methods that perform camera pose refinement from a set of inaccurate initial poses.