simple-romp 0.0.4

ROMP: Monocular, One-stage, Regression of Multiple 3D People, ICCV21

Simple_ROMP

Simplified implementation of ROMP: Monocular, One-stage, Regression of Multiple 3D People, ICCV21

Please refer to https://github.com/Arthur151/ROMP for more details

Installation

pip install --upgrade setuptools numpy cython

pip install simple_romp

or download the package and install it from source:

python setup.py install

Usage

Webcam demo:

romp --mode=webcam --show

For Mac Users, please use the original terminal instead of other terminal app (e.g. iTerm2) to avoid the bug zsh: abort.

Processing a single image:

romp --mode=image --calc_smpl --render_mesh --input=/path/to/image.jpg --save_path=/path/to/results.jpg

Processing a folder of images:

romp --mode=video --calc_smpl --render_mesh  --input=/path/to/image/folder/ --save_path=/path/to/output/folder/

Processing a video:

romp --mode=video --calc_smpl --render_mesh  --input=/path/to/video.mp4 --save_path=/path/to/output/folder/results.mp4 --save_video

Optional functions:

# show the results during processing image / video, add:
--show

# to smooth the results in webcam / video processing, add: (the smaller the smooth_coeff, the smoother) 
--temporal_optimize --smooth_coeff=3.

# to use the onnx version of ROMP for faster inference, please add:
--onnx

# to show the largest person only (remove the small subjects in background), add:
--show_largest 

More options, see romp -h

Note that if you are using CPU for inference, we highly recommand to add --onnx for much faster speed.

Tools

To convert the trained ROMP model '.pkl' (like ROMP.pkl) to simple-romp '.pth' model, please run

cd /path/to/ROMP/simple_romp/
python tools/convert_checkpoints.py ROMP.pkl ROMP.pth

Copyright

Codes released under MIT license. All rights reserved by Yu Sun.