U2-Net: Going deeper with nested U-structure for salient object detection

doi:10.1016/j.patcog.2020.107404

U2-Net: Going deeper with nested U-structure for salient object detection

2020-10-01

1.2K

OA

AI

DOI：10.1016/j.patcog.2020.107404

原文链接

查看原文

摘要

En 中文

In this paper, we design a simple yet powerful deep network architecture, U-2-Net, for salient object detection (SOD). The architecture of our U-2-Net is a two-level nested U-structure. The design has the following advantages: (1) it is able to capture more contextual information from different scales thanks to the mixture of receptive fields of different sizes in our proposed ReSidual U-blocks (RSU), (2) it increases the depth of the whole architecture without significantly increasing the computational cost because of the pooling operations used in these RSU blocks. This architecture enables us to train a deep network from scratch without using backbones from image classification tasks. We instantiate two models of the proposed architecture, U-2-Net (176.3 MB, 30 FPS on GTX 1080Ti GPU) and U-2-Nett (4.7 MB, 40 FPS), to facilitate the usage in different environments. Both models achieve competitive performance on six SOD datasets. The code is available: https://github.com/NathanUA/U-2-Net. (C) 2020 Elsevier Ltd. All rights reserved.

Keyword:

Salient object detection

Convolutional neural network

Network architecture design

Nested U-structure

Multi-scale feature extraction

U2-Net: Going deeper with nested U-structure for salient object detection

摘要

AI总结

期刊

机构

引用论文