search

hongsukchoi / Pose2Mesh_RELEASE

Official Pytorch implementation of "Pose2Mesh: Graph Convolutional Network for 3D Human Pose and Mesh Recovery from a 2D Human Pose", ECCV 2020
MIT License
600 stars 69 forks source link

image annotation tool #45

Open bsraina opened 2 years ago

bsraina commented 2 years ago

hello dear can you tell me which annotation tool you have use for annotating images. and after which i have to run , to create .npy file and also can you give step by step tutorial

thanx in advance

bsraina commented 2 years ago

i want to create this type of 3d model can i get using your repositorie?

hongsukchoi commented 2 years ago

hello dear can you tell me which annotation tool you have use for annotating images. and after which i have to run , to create .npy file and also can you give step by step tutorial

thanx in advance

I don't underrstand. I did not annotate any images myself. Are you trying to make a new dataset?

i want to create this type of 3d model can i get using your repositorie?

You can get the obj result file by running demo code.

bsraina commented 2 years ago

no i want to make .npy of my own pictures or demo_input_pose2d.json file to get my own model, please help me dear thank you again in advance.

vfr-xcugas commented 2 years ago

can you make a tutorial video on it? I will be grateful to you thanx in advance.🙏

hongsukchoi commented 2 years ago

no i want to make .npy of my own pictures or demo_input_pose2d.json file to get my own model, please help me dear thank you again in advance.

I see. You have to use off-the-shelf 2D pose detectors to get the input 2D pose for Pose2Mesh. Check the demo codes of them: https://github.com/leoxiaobin/deep-high-resolution-net.pytorch/tree/master/demo https://github.com/tensorboy/pytorch_Realtime_Multi-Person_Pose_Estimation https://github.com/HRNet/HigherHRNet-Human-Pose-Estimation

Run their demo codes on your own image.

can you make a tutorial video on it? I will be grateful to you thanx in advance.🙏

Thanks for your interest, but making a tutorial is too much burden to me. If you think obtaining 2D pose in advance is cumbersome, check my another work, which only requires images as input; https://github.com/hongsukchoi/TCMR_RELEASE