RANet: Ranking Attention Network for Fast Video Object Segmentation

Ziqin Wang "RANet: Ranking Attention Network for Fast Video Object Segmentation", ICCV 2019, official version, arXiv

---

Contact Information

Ziqin Wang

Contents

1. Introduction
2. Code
3. Download
4. Others
5. Citation

Introduction

1. Overview

<img src="pics/overview.png" width="400px"/>

2. Framework

<img src="pics/RANet.png" width="800px"/>

3. Ranking

<img src="pics/Ranking.png" width="400px"/>

Code

1. Requirement

Pytorch (tested on 1.0.1 and 0.4.1)

torchvision = 0.2

2. Usage

1. Download the pretained model from this page.
2. Link DAVIS folder into datasets folder. (Please download DAVIS 2017 version.)
3. Run RANet.py

Download

Paper

Supplementary File

Precomputed results:Google drive

Pretrained models:Baidu, Google drive