Jetson-inference posenet python command help

Hi,

I’m wanting to change the colors of each keypoint detected by the posenet model individually. I looked at the c++ source code and there appears to be a command to do this.

To my knowledge, the python command binding doesn’t pass this command through. There is a python command to change the opacity and I’ve been playing around with this command to test the concept out.

It works when I change every keypoints’ opacity, but causes an error when I try to change just one keypoint’s opacity.


My code is based on an example provided in the jetson-inference repository…

#!/usr/bin/python3
#
# Copyright (c) 2021, NVIDIA CORPORATION. All rights reserved.
#
# Permission is hereby granted, free of charge, to any person obtaining a
# copy of this software and associated documentation files (the "Software"),
# to deal in the Software without restriction, including without limitation
# the rights to use, copy, modify, merge, publish, distribute, sublicense,
# and/or sell copies of the Software, and to permit persons to whom the
# Software is furnished to do so, subject to the following conditions:
#
# The above copyright notice and this permission notice shall be included in
# all copies or substantial portions of the Software.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
# THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
# FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
# DEALINGS IN THE SOFTWARE.
#

import jetson.inference
import jetson.utils

import argparse
import sys

# parse the command line
parser = argparse.ArgumentParser(description="Run pose estimation DNN on a video/image stream.", 
                                 formatter_class=argparse.RawTextHelpFormatter, epilog=jetson.inference.poseNet.Usage() +
                                 jetson.utils.videoSource.Usage() + jetson.utils.videoOutput.Usage() + jetson.utils.logUsage())

parser.add_argument("input_URI", type=str, default="", nargs='?', help="URI of the input stream")
parser.add_argument("output_URI", type=str, default="", nargs='?', help="URI of the output stream")
parser.add_argument("--network", type=str, default="resnet18-body", help="pre-trained model to load (see below for options)")
parser.add_argument("--overlay", type=str, default="links,keypoints", help="pose overlay flags (e.g. --overlay=links,keypoints)\nvalid combinations are:  'links', 'keypoints', 'boxes', 'none'")
parser.add_argument("--threshold", type=float, default=0.15, help="minimum detection threshold to use") 

try:
	opt = parser.parse_known_args()[0]
except:
	print("")
	parser.print_help()
	sys.exit(0)

# load the pose estimation model
net = jetson.inference.poseNet(opt.network, sys.argv, opt.threshold)

# create video sources & outputs
input = jetson.utils.videoSource(opt.input_URI, argv=sys.argv)
output = jetson.utils.videoOutput(opt.output_URI, argv=sys.argv)

# process frames until the user exits
while True:
    # capture the next image
    img = input.Capture()

    # perform pose estimation (with overlay)
    poses = net.Process(img, overlay=opt.overlay)

    # print the pose results
    # print("detected {:d} objects in image".format(len(poses)))
    # [TRT]    poseNet -- keypoint 00 'nose'  color 0 255 0 255
    # [TRT]    poseNet -- keypoint 01 'left_eye'  color 0 255 0 255
    # [TRT]    poseNet -- keypoint 02 'right_eye'  color 0 255 0 255
    # [TRT]    poseNet -- keypoint 03 'left_ear'  color 0 255 0 255
    # [TRT]    poseNet -- keypoint 04 'right_ear'  color 0 255 0 255
    # [TRT]    poseNet -- keypoint 05 'left_shoulder'  color 0 255 0 255
    # [TRT]    poseNet -- keypoint 06 'right_shoulder'  color 0 255 0 255
    # [TRT]    poseNet -- keypoint 07 'left_elbow'  color 0 255 0 255
    # [TRT]    poseNet -- keypoint 08 'right_elbow'  color 0 255 0 255
    # [TRT]    poseNet -- keypoint 09 'left_wrist'  color 0 255 0 255
    # [TRT]    poseNet -- keypoint 10 'right_wrist'  color 0 255 0 255
    # [TRT]    poseNet -- keypoint 11 'left_hip'  color 0 255 0 255
    # [TRT]    poseNet -- keypoint 12 'right_hip'  color 0 255 0 255
    # [TRT]    poseNet -- keypoint 13 'left_knee'  color 0 255 0 255
    # [TRT]    poseNet -- keypoint 14 'right_knee'  color 0 255 0 255
    # [TRT]    poseNet -- keypoint 15 'left_ankle'  color 0 255 0 255
    # [TRT]    poseNet -- keypoint 16 'right_ankle'  color 0 255 0 255
    # [TRT]    poseNet -- keypoint 17 'neck'  color 0 255 0 255
    # [TRT]    poseNet -- loaded 18 class colors 
    # index 6 is right shoulder
    # Code below causes error
    # net.SetKeypointAlpha(alpha=20.0,keypoint=6) 
    for pose in poses:
        print(pose)
        print(pose.Keypoints)
        print(net.GetNumKeypoints())
        print('Links', pose.Links)

    # render the image
    output.Render(img)
    
    # update the title bar
    output.SetStatus("{:s} | Network {:.0f} FPS".format(opt.network, net.GetNetworkFPS()))

    # print out performance info
    # net.PrintProfilerTimes()

    # exit on input/output EOS
    if not input.IsStreaming() or not output.IsStreaming():
        break

I’m currently going to try to recode the example in c++ and see if that works. Any advise is appreciated.

Thanks

Hi @alexr1111111, there is a colors text file that comes with each model that you can use to persistently change the colors if that’s easier. These are specified one color per line, R G B A (so you can specify the opacity there too)

The order of the colors in the text file are the same as the order of the keypoints. You are correct that some of the Python bindings are missing, sorry about that!

So, RGB color values in the color.txt file is checked continuously while the posenet model is running? Meaning, I would not have to restart the script to in order for the new colors to be instantiated? I’m trying to create a dynamic pressure map using color and the posenet outline. In this application, the color values could be changing every couple of seconds. Anyhow, I’ll test it out. Thanks!

I see - no, they are loaded just once. In your case with dynamic colors, you are correct that you would want to use the poseNet::SetKeypointColor() function. You can add a binding to it from Python in jetson-inference/python/bindings/PyPoseNet.cpp (sorry I will make a note these are missing currently)

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.