ROMP: Monocular, One-stage, Regression of Multiple 3D People, ICCV21
Project description
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 black
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 show the largest person only (remove the small subjects in background), add:
--show_largest
More options, see romp -h
Copyright
Codes released under MIT license. All rights reserved by Yu Sun.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
simple_romp-0.0.3.tar.gz
(74.2 kB
view hashes)
Built Distribution
Close
Hashes for simple_romp-0.0.3-cp38-cp38-manylinux1_x86_64.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | f3aba9cf439e6a058086ce0690490e2f7c20aba76820628bf729632d6b6a8a2f |
|
MD5 | f76081728350b20be7b688f1262931a8 |
|
BLAKE2b-256 | 4f44044a2935847e74d1a869df7bb76c3f1da7aac7d4f952443b6b4494d0aabf |