Robust-CVD

[Davidyao99]

I notice that you ran Robust Consistent Video Depth as one of your baselines. I have been able to run their codebase found here.

May I ask how did you manage to retrieve the pose from RCVD? It seems like the default settings only return depth. The codebase does not seem to be maintained anymore. Wondering if you could provide some guidance, that will be very helpful! Thanks so much!

[Junyi42]

Hi @Davidyao99,

Could you please refer to facebookresearch/robust_cvd#6 (specifically, @ztzhang's reply) to get the camera pose?

Hope this helps!

[Davidyao99]

Thanks a lot I must have missed that issue! When running RCVD, did you use any separate hyperparameters? or just stick with the default ones as shown in their colab (with frame_range adjusted to cover the entire video)