search

stefanopini / simple-HRNet

Multi-person Human Pose Estimation with HRNet in Pytorch
GNU General Public License v3.0
579 stars 108 forks source link

What is the meaning of each column in generated CSV file? #82

Closed jellyfish1456 closed 3 years ago

jellyfish1456 commented 3 years ago

Hi~ I have tried python scripts/extract-keypoints.py --filename input.mp4 --format csv I successfully get the csv file. However, I cannot understand the column of the coordinate.

Could you please list the meaning of each column.

Thank you very much!

image

stefanopini commented 3 years ago

Hi!

It is reported here https://github.com/stefanopini/simple-HRNet/blob/master/scripts/extract-keypoints.py#L81-L84

csv format is:
   frame_index,detection_index,<point 0>,<point 1>,...,<point hrnet_j>
 where each <point N> corresponds to three elements:
   y_coordinate,x_coordinate,confidence

So each row contains: Frame id - detection id - keypoint 0, y coordinate - keypoint 0, x coordinate, keypoint 0, confidence - keypoint 1 ...

Let me know if you need any other question!