preface

In the year of 2020, this article once again updates the latest semantic segmentation papers worth attention. This paper will be pushed to Github, welcome everyone star/fork (click to read the original article, or directly access) :

Github.com/amusi/daily…

Matters needing attention:

  • It includes both aerial image semantic segmentation network and domain adaptive open source network
  • Publication period: January 03, 2020 – January 29, 2020

Semantic segmentation thesis

[1] Graph-FCN: Graph convolutional network for image semantic segmentation

Graph-fcn for Image Semantic segmentation

Time: 20200103

Author team: Chinese Academy of Sciences & UcAS & Beijing University of Chinese Medicine

Link: arxiv.org/abs/2001.00…

Note: As far As we know, it is the first time that we apply the graph convolutional network in image semantic segmentation

Graph-FCN\

[2] ExtremeC3Net: Extreme lightweight portrait segmentation network (using advanced C3 module)

ExtremeC3Net: Extreme Lightweight Portrait Segmentation Networks Using Advanced C3-Modules

Time: 20190812

Author team: Seoul National University &Clova AI

Link: arxiv.org/abs/1908.03…

Code: github.com/clovaai/ext…

Note 1: This article has been pushed by CVer before, but it was not open source at that time. Not long ago, ExtremeC3Net has just opened source, so I share it with you again, it is worth paying attention and learning!

Note 2: Only 37.7K parameters! Better performance than PortraitNet, BiSeNet and ESPNetV2 networks, code and data set is now open source!

[3] HMANet: A hybrid multi-attention network for aerial image semantic segmentation

HMANet: Hybrid Multiple Attention Network for Semantic Segmentation in Aerial Images

Time: 20200110

Author team: Chinese Academy of Sciences & UcAS

Link: arxiv.org/abs/2001.02…

Note: On Vaihingen/Potsdam and other datasets, the performance of SOTA! The performance is better than CCNet, ACFNet, PSPNet and DeepLabV3+

[4] EVS: Efficient video semantic segmentation with tag propagation and optimization functions

“Efficient Video Semantic Segmentation with Labels Propagation and Refinement”

Time: 20200112 (WACV2020)

Team of authors: CV Laboratory, ETH Zurich

Link: arxiv.org/abs/1912.11…

Note: On Cityscapes (2048 x 1024), speeds are 80-1000 FPS! The performance is better than DVSN, ICNet and other networks

[5] Unsupervised domain adaptive mobile semantic segmentation based on cyclic consistency and feature alignment

Unsupervised Domain Adaptation for Mobile Semantic Segmentation Based on Cycle Consistency and Feature Alignment

Time: 20200115

Team of authors: University of Padua

Link: arxiv.org/abs/2001.04…

Note: The performance is better than CycleGAN and CyCADA networks

[6] GPSNet: Gated path selective semantic segmentation network

“Gated Path Selection Network for Semantic Segmentation”

Time: 20200122

Authors: Beihang university, Oxford University, Tsinghua University, etc

Link: arxiv.org/abs/2001.06…

Note: The performance is better than OCNet, PSPNet and PSANet networks

GPSNet

[7] DADA: Depth perception domain adaptation for semantic segmentation

DADA: Depth-Aware Domain Adaptation in Semantic Segmentation

Time: 20190426 (ICCV 2019)

Team of authors: Valeo AI Lab& Sorbonne University

Link: arxiv.org/abs/1904.01…

Code: github.com/valeoai/DAD…

Note 1: This article was previously pushed by CVer, but it was not open source at that time. DADA just opened source, so it is again shared to everyone, worth paying attention and learning!

Note 2: Performance better than AdvEnt, AdaptPatch and CLAN networks, performance SOTA, now open source!

[8] CAG-UDA: category Anchor guidance for adaptive semantic segmentation in unsupervised domain

Category Anchor-Guided Unsupervised Domain Adaptation for Semantic Segmentation

Time: 20191017

Author team: University of Sydney & Tencent Youtu

Link: arxiv.org/pdf/1910.13…

Code: github.com/RogerZhangz…

Note 1: This article has been pushed by CVer before, but it was not open source at that time, not long ago, CAG-UDA just open source, so again share to everyone, worth paying attention to and learning!

Note 2: Performance is better than DCAN, CLAN and BLF networks

CAG-UDA

[9] Open-set semantic segmentation network for aerial remote sensing images

“Towards Open-Set Semantic Segmentation of Aerial Images”

Time: 20200129

Author team: Stirling University, Et al

Link: arxiv.org/abs/2001.10…

Note: According to the author, this is the first paper on semantic segmentation technology applied to open set scenes of remote sensing images (Open set refers to the existence of unknown classes).

In order to facilitate download, I have packaged the above paper, and you can get the package link by replying to the background of CVer official account: 20200220.

“`php

Highlights of past For beginners entry route of artificial intelligence and data download AI based machine learning online manual deep learning online manual download update (PDF to 25 sets) note: WeChat group or qq group to join this site, please reply “add group” to get a sale standing knowledge star coupons, please reply “planet” knowledge like articles, point in watching

Copy the code