Home

Awesome

Weakly Supervised Learning of Instance Segmentation with Inter-pixel Relations

<p align="center"><img src="outline.jpg" alt="outline" width="90%"></p> The code of:

Weakly Supervised Learning of Instance Segmentation with Inter-pixel Relations, Jiwoon Ahn, Sunghyun Cho, and Suha Kwak, CVPR 2019 [Paper]

This repository contains a framework for learning instance segmentation with image-level class labels as supervision. The key component of our approach is Inter-pixel Relation Network (IRNet) that estimates two types of information: a displacement vector field and a class boundary map, both of which are in turn used to generate pseudo instance masks from CAMs.

Citation

If you find the code useful, please consider citing our paper using the following BibTeX entry.

@InProceedings{Ahn_2019_CVPR,
author = {Ahn, Jiwoon and Cho, Sunghyun and Kwak, Suha},
title = {Weakly Supervised Learning of Instance Segmentation with Inter-pixel Relations},
booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
month = {June},
year = {2019}
}

Prerequisite

Usage

Install python dependencies

pip install -r requirements.txt

Download PASCAL VOC 2012 devkit

Run run_sample.py or make your own script

python run_sample.py

Train Mask R-CNN or DeepLab with the generated pseudo labels

TO DO