Giter Club home page Giter Club logo

Comments (7)

RiRyuichi avatar RiRyuichi commented on August 20, 2024 1

Dear Rajivo,

Thank you for your information.
I edited Person.msg, rosOpenposeAsync.cpp and rosOpenposeSync.cpp files, and run with
text = [facePart.pixel for person in msg.persons for facePart in person.faceParts] in run.lauch.
Now I can output coordinates of face keypoints as shown below-

[INFO] [1667808734.583089]: [x: 309.166168213
y: 184.518478394, x: 311.559112549
y: 203.063949585, x: 315.14855957
y: 218.019973755, x: 320.532745361
y: 233.574249268, x: 328.309875488
y: 246.735549927, x: 338.479980469
y: 260.49508667, x: 349.846557617
y: 270.66519165, x: 364.204315186
y: 279.638793945, x: 379.758575439
y: 280.237030029, x: 393.518127441
y: 275.45111084, x: 403.688232422
y: 265.879241943, x: 412.063598633
y: 255.110916138, x: 419.84072876
y: 241.949615479, x: 424.028411865
y: 228.788314819, x: 427.019622803
y: 215.62701416, x: 428.216094971
y: 200.67098999, x: 427.019622803
y: 186.313201904, x: 331.301086426
y: 164.776535034, x: 340.872924805
y: 158.794128418, x: 353.435974121
y: 155.80291748, x: 364.204315186
y: 159.392364502, x: 374.97265625
y: 164.17829895, x: 388.732208252
y: 164.17829895, x: 397.705810547
y: 158.794128418, x: 407.277679443
y: 155.80291748, x: 416.849517822
y: 157.597640991, x: 421.635467529
y: 165.374771118, x: 382.749786377
y: 179.13432312, x: 382.749786377
y: 186.911453247, x: 382.749786377
y: 194.09034729, x: 383.348022461
y: 203.063949585, x: 371.981445312
y: 217.421737671, x: 377.365631104
y: 218.019973755, x: 382.749786377
y: 218.618225098, x: 387.535736084
y: 218.019973755, x: 392.321655273
y: 217.421737671, x: 344.462371826
y: 180.330795288, x: 350.444793701
y: 179.732559204, x: 358.820159912
y: 179.732559204, x: 364.204315186
y: 183.92024231, x: 358.820159912
y: 183.92024231, x: 349.846557617
y: 183.92024231, x: 396.509338379
y: 183.322006226, x: 402.491760254
y: 179.732559204, x: 408.474151611
y: 179.732559204, x: 415.05480957
y: 180.929046631, x: 409.072387695
y: 183.92024231, x: 402.491760254
y: 183.92024231, x: 363.606079102
y: 241.351379395, x: 369.588500977
y: 232.975997925, x: 377.963867188
y: 228.190078735, x: 382.749786377
y: 230.58303833, x: 388.133972168
y: 228.190078735, x: 394.116363525
y: 232.975997925, x: 398.304046631
y: 241.351379395, x: 393.518127441
y: 246.735549927, x: 388.133972168
y: 250.324981689, x: 382.749786377
y: 250.324981689, x: 377.963867188
y: 250.324981689, x: 368.990264893
y: 246.735549927, x: 368.391998291
y: 241.351379395, x: 377.963867188
y: 237.761932373, x: 382.749786377
y: 237.761932373, x: 387.535736084
y: 237.761932373, x: 395.911102295
y: 240.753128052, x: 387.535736084
y: 237.761932373, x: 382.749786377
y: 238.958404541, x: 377.963867188
y: 238.360168457, x: 354.034240723
y: 180.929046631, x: 405.482940674
y: 180.330795288]

The accuracy of the numerical value has not been confirmed yet, but I think the mapping of face key points in the topic will just like be official format.

Thank you very much! :-)

from ros_openpose.

ravijo avatar ravijo commented on August 20, 2024 1

Thanks @RiRyuichi

It is perfect 👍

The pull request is merged. afecab6

Thanks again for the great contribution.

from ros_openpose.

ravijo avatar ravijo commented on August 20, 2024

@RiRyuuichi

I am glad to see your message. Thank you.

Yes, you are right. You need to edit the following line:

ros_openpose/scripts/echo.py

Line 14 in c6f5ce7

text = [bodyPart.pixel for person in msg.persons for bodyPart in person.bodyParts]

However, first, you should check the /frame topic to confirm that face key points are there. I have not used it with face key points, but some users did in the past.

Next, you should figure out the mapping of face key points in the topic. For example, please see the below mapping for complete body tracking:

ros_openpose/scripts/visualizer.py

Lines 43 to 60 in c6f5ce7

See the link below to get the id of each joint as defined in Kinect v2
src: https://github.com/CMU-Perceptual-Computing-Lab/openpose/blob/master/doc/output.md#keypoint-ordering
Result for BODY_25 (25 body parts consisting of COCO + foot)
const std::map<unsigned int, std::string> POSE_BODY_25_BODY_PARTS {
{ 0, "Nose"}, {13, "LKnee"}
{ 1, "Neck"}, {14, "LAnkle"}
{ 2, "RShoulder"}, {15, "REye"}
{ 3, "RElbow"}, {16, "LEye"}
{ 4, "RWrist"}, {17, "REar"}
{ 5, "LShoulder"}, {18, "LEar"}
{ 6, "LElbow"}, {19, "LBigToe"}
{ 7, "LWrist"}, {20, "LSmallToe"}
{ 8, "MidHip"}, {21, "LHeel"}
{ 9, "RHip"}, {22, "RBigToe"}
{10, "RKnee"}, {23, "RSmallToe"}
{11, "RAnkle"}, {24, "RHeel"}
{12, "LHip"}, {25, "Background"}

Finally, you can use conditional statement to print only face key points based on the index.

Let me know how it goes!

Cheers.

from ros_openpose.

RiRyuichi avatar RiRyuichi commented on August 20, 2024

Dear Rajivo,

Thank you for the prompt reply.
I checked the /frame topic and it repeatedly showed :

header: 
  seq: 69
  stamp: 
    secs: 1667797310
    nsecs: 784934479
  frame_id: "camera_color_optical_frame"
persons: 
  - 
    bodyParts: 
      - 
        score: 0.79394453764
        pixel: 
          x: 444.27545166
          y: 210.642166138
        point: 
          x: 0.121169157326
          y: -0.0344984829426
          z: 0.658000051975
      - 
        score: 0.56282800436
        pixel: 
          x: 403.805603027
          y: 326.811676025
        point: 
          x: 0.0890130102634
          y: 0.102006264031
          z: 0.75100004673
      - 
        score: 0.327622026205
        pixel: 
          x: 262.836120605
          y: 341.170684814
        point: 
          x: -0.0823214277625
          y: 0.119004175067
          z: 0.748000025749
      - 
        score: 0.0
        pixel: 
          x: 0.0
          y: 0.0
        point: 
          x: -1.91599595547
          y: -1.40698289871
          z: 3.57300019264
      - 
        score: 0.0
        pixel: 
          x: 0.0
          y: 0.0
        point: 
          x: -1.91599595547
          y: -1.40698289871
          z: 3.57300019264
      - 
        score: 0.444403111935
        pixel: 
          x: 540.85559082
          y: 308.534851074
        point: 
          x: 0.282483309507
          y: 0.0880472958088
          z: 0.82900005579
      - 
        score: 0.237627774477
        pixel: 
          x: 590.471923828
          y: 454.699707031
        point: 
          x: 0.312954962254
          y: 0.254904001951
          z: 0.743000030518
      - 
        score: 0.0
        pixel: 
          x: 0.0
          y: 0.0
        point: 
          x: -1.91599595547
          y: -1.40698289871
          z: 3.57300019264
      - 
        score: 0.0
        pixel: 
          x: 0.0
          y: 0.0
        point: 
          x: -1.91599595547
          y: -1.40698289871
          z: 3.57300019264
      - 
        score: 0.0
        pixel: 
          x: 0.0
          y: 0.0
        point: 
          x: -1.91599595547
          y: -1.40698289871
          z: 3.57300019264
      - 
        score: 0.0
        pixel: 
          x: 0.0
          y: 0.0
        point: 
          x: -1.91599595547
          y: -1.40698289871
          z: 3.57300019264
      - 
        score: 0.0
        pixel: 
          x: 0.0
          y: 0.0
        point: 
          x: -1.91599595547
          y: -1.40698289871
          z: 3.57300019264
      - 
        score: 0.0
        pixel: 
          x: 0.0
          y: 0.0
        point: 
          x: -1.91599595547
          y: -1.40698289871
          z: 3.57300019264
      - 
        score: 0.0
        pixel: 
          x: 0.0
          y: 0.0
        point: 
          x: -1.91599595547
          y: -1.40698289871
          z: 3.57300019264
      - 
        score: 0.0
        pixel: 
          x: 0.0
          y: 0.0
        point: 
          x: -1.91599595547
          y: -1.40698289871
          z: 3.57300019264
      - 
        score: 0.848506808281
        pixel: 
          x: 412.987182617
          y: 180.612442017
        point: 
          x: 0.159630015492
          y: -0.0722588449717
          z: 0.686000049114
      - 
        score: 0.839394986629
        pixel: 
          x: 474.21484375
          y: 175.371795654
        point: 
          x: 0.159628242254
          y: -0.0751760751009
          z: 0.686000049114
      - 
        score: 0.814783334732
        pixel: 
          x: 350.308013916
          y: 187.127853394
        point: 
          x: 0.0234853327274
          y: -0.0669053867459
          z: 0.739000022411
      - 
        score: 0.484164863825
        pixel: 
          x: 497.76574707
          y: 183.194992065
        point: 
          x: 0.185554116964
          y: -0.0663822293282
          z: 0.685000061989
      - 
        score: 0.0
        pixel: 
          x: 0.0
          y: 0.0
        point: 
          x: -1.91599595547
          y: -1.40698289871
          z: 3.57300019264
      - 
        score: 0.0
        pixel: 
          x: 0.0
          y: 0.0
        point: 
          x: -1.91599595547
          y: -1.40698289871
          z: 3.57300019264
      - 
        score: 0.0
        pixel: 
          x: 0.0
          y: 0.0
        point: 
          x: -1.91599595547
          y: -1.40698289871
          z: 3.57300019264
      - 
        score: 0.0
        pixel: 
          x: 0.0
          y: 0.0
        point: 
          x: -1.91599595547
          y: -1.40698289871
          z: 3.57300019264
      - 
        score: 0.0
        pixel: 
          x: 0.0
          y: 0.0
        point: 
          x: -1.91599595547
          y: -1.40698289871
          z: 3.57300019264
      - 
        score: 0.0
        pixel: 
          x: 0.0
          y: 0.0
        point: 
          x: -1.91599595547
          y: -1.40698289871
          z: 3.57300019264
    leftHandParts: []
    rightHandParts: []
---

It looks like only showed bodyParts key points, and support for leftHandParts and rightHandParts. But not support for face key points...
Also I checked visualizer.py file, there are some contents about bodyParts, leftHandParts and rightHandParts, but nothing about face parts.
Should I add something in visualizer.py file?

from ros_openpose.

ravijo avatar ravijo commented on August 20, 2024

@RiRyuuichi

Thank you very much. As I said, I never used --face flag. Therefore, it could be possible that it is not implemented. But some users have used it previously for face as well.

Please see Person definition, i.e., Person.msg below:

ros_openpose/msg/Person.msg

Lines 1 to 5 in c6f5ce7

# A person has some body parts. That is why we have created
# an array of body parts.
BodyPart[] bodyParts
BodyPart[] leftHandParts
BodyPart[] rightHandParts

You should add one more attribute say BodyPart[] faceParts and then fill this field. You can easily take help from the existing code. Please see how leftHandParts and rightHandParts are filled.

Let me know your progress.

from ros_openpose.

ravijo avatar ravijo commented on August 20, 2024

@RiRyuuichi

I am glad to see your speedy progress. Well done!!!

May I request you create a pull request to share your code? In this way, we can merge it with this repository. This helpful information can be made available to others.

Cheers!

from ros_openpose.

RiRyuichi avatar RiRyuichi commented on August 20, 2024

Dear Rajivo,

I tried to create a pull request to share codes. I'm sorry that I'm a github beginner...
Is this alright?

from ros_openpose.

Related Issues (20)

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.