Tf-pose-estimation: The run_video.py my be wrong in Ubuntu16.04

[juciny]

when I run run_video.py, Bone spots are not detected
so I compare run_video.py with run.py
humans = e.inference(image,resize_to_default=(w > 0 and h > 0), upsample_size=args.resize_out_ratio)
that is right
maybe my environment configuration is wrong

[JakubWojciechowski]

Hey,
for me, run_video.py did not work properly with the default upsample_size too. After changing this parameter to the value from run.py:
humans = e.inference (image, upsample_size = 4.0)
everything works correctly.
Good luck!

[Aditya-11]

Thank you very much @JakubWojciechowski

[dlwogns0128]

Thank you so much @JakubWojciechowski :0