ERSOD-Net

by Xiurong Jiang, Lin Zhu, Hui Tian

Motivation

The motivation and visual results of our method.

Framework

The pipeline of the proposed approach.

LSTM-based module

Overview of LSTM-based module.

Prerequisites

• Python 3.5
• Pytorch 1.3
• OpenCV 4.0
• Numpy 1.15
• TensorboardX

Download dataset

Download the following datasets and unzip them into data folder

• DUTS
• DUT-OMRON
• ECSSD
• HKU-IS
• PASCAL-S
• ERSOD

Saliency maps & Trained model

• saliency maps: Google

• trained model: Google

• The simulated event data and label

• Quantitative comparisons on five SOD datasets

• Qualitative comparisons with state-of-the-arts methods

• Qualitative comparisons on real Event-RGB SOD dataset