demo.py result explanation

[Hunamic]

Hello,

Thank you for your excellent work on this project!

I have a question regarding the outputs generated by running the demo.py file. Specifically, I want to understand the rows in result/images.txt and saved_trajectory/results.txt. From what I gather, they seem to represent camera poses relative to the world frame.

However, when I test with the TartanAir dataset and the Seasonsforest_winter_sample_P006 sequence, which contains 330 frames, the model outputs only 165 rows for the poses—exactly half of the total video frames(and also the poses related to the world frame).

Could you please clarify the following:

1. What do the rows in these files represent?
2. Why does the model output only half the number of poses relative to the frames in the sequence?
   Apologies if this question is overly simplistic, and I appreciate your response in advance.

Thank you!

[herrouou]

Hi, how did you get the final result? My running seemed to get stuck there:
(poses, tstamps), (points, colors, calib) = run(cfg, args.network, args.imagedir, args.calib, args.stride, args.skip, args.viz, args.timeit)
Some return values didn't work well....