/Mask-Selection-Networks

[CVPR 2021] Youtube-VIS 2021 3rd place, [CVPR 2020] winner DAVIS 2020. Code for mask selection based methods.

Mask-Selection-Networks

[CVPRW 2021] [WACV] [CVPRW 2020]

This is the repo to host the code for mask selection based methods of video instance segmentation and the related problem of unsupervised video object segmentation. The method won various challenges such as DAVIS 2020, Youtube VIS 2021. The code for our WACV paper can be found here.

vis

Introduction

In this work we present a novel solution for Video Instance Segmentation(VIS), that is automatically generating instance level segmentation masks along with object class and tracking them in a video. Our method improves the masks from segmentation and propagation branches in an online manner using the Mask Selection Network (MSN) hence limiting the noise accumulation during mask tracking. We propose an effective design of MSN by using patch-based convolutional neural network. The network is able to distinguish between very subtle differences between the masks and choose the better masks out of the associated masks accurately. Further we make use of temporal consistency and process the video sequences in both forward and reverse manner as a post processing step to recover lost objects. The proposed method can be used to adapt any video object segmentation method for the task of VIS. Our method achieves a score of 49.1 mAP on 2021 YouTube-VIS Challenge and was ranked third place among more than 30 global teams.

Method Overview

overview

Results

results

Bibtex

Please cite the following papers if the work was helpful.

@article{goel2021msn,
  title={MSN: Efficient Online Mask Selection Network for Video Instance Segmentation},
  author={Goel, Vidit and Li, Jiachen and Garg, Shubhika and Maheshwari, Harsh and Shi, Humphrey},
  journal={arXiv preprint arXiv:2106.10452},
  year={2021}
}
@inproceedings{garg2021mask,
  title={Mask Selection and Propagation for Unsupervised Video Object Segmentation},
  author={Garg, Shubhika and Goel, Vidit},
  booktitle={Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision},
  pages={1680--1690},
  year={2021}
}
@article{DAVIS2020-Unsupervised-1st,
              author = {S. Garg, V. Goel, S. Kumar},
              title = {Unsupervised Video Object Segmentation using Online Mask Selection and Space-time Memory Networks},
              journal = {The 2020 DAVIS Challenge on Video Object Segmentation - CVPR Workshops},
              year = {2020}
}