Training Code for IPN
This repository implemented training code for "Fast User-Guided Video Object Segmentation by Interaction-and-Propagation Networks" (CVPR 2019) based on its official demo code.
Create the environment
# create conda env
conda create -n ipn python=3.7
# activate conda env
conda activate ipn
# install pytorch
conda install pytorch=1.3 torchvision
# install other dependencies
pip install -r requirements.txt
Dataset Preparation
- DAVIS 2017 Dataset
- Download the data and human annotated scribbles here.
- Place
DAVIS
folder intoroot/data
.
Training
python ipn_trainval.py
trained models will be stored in root/weights
. You can download the trained models here.
Evaluation
The evaluation code can be found here.
Reference
Fast User-Guided Video Object Segmentation by Interaction-and-Propagation Networks
Seoung Wug Oh, Joon-Young Lee, Ning Xu, Seon Joo Kim
CVPR 2019