No pose output about realtime_multi-person_pose_estimation HOT 2 CLOSED

This problem is mainly caused by openCV version difference as reported in CMU-Perceptual-Computing-Lab/caffe_rtpose#2 and CMU-Perceptual-Computing-Lab/caffe_rtpose#5. We made a commit in e1388f7 Did you use the updated version? One way to debug the issue is to turn off the display and write out the result, e.g., ./build/examples/rtpose/rtpose.bin --video test.mp4 --no_display --write_frames output/ Please let us know if you meet the same problem

from realtime_multi-person_pose_estimation.

Hi again,
thank you for your response and sorry for the long pause, I was trying different things to solve this issue based on your feedback.
I have tried various things in the meantime and fortunately today I have finally got it to work!
I am not sure what exactly was the problem but, for the record, I write here my current (working) configuration.
I now use the latest commits of caffe_rtpose (122d9db) and Realtime_Multi-Person_Pose_Estimation
(4fa16c5), with CUDA 8.0, cuDNN 5.1.5, OpenCV 2.4.12 and NVIDIA graphics driver version 375.39.
Now it works perfectly also with the display turned on.
Thank you again for your feedback!

from realtime_multi-person_pose_estimation.