語義分割paper02

轉自：   http://blog.csdn.net/fabulousli/article/details/78633531

在FCN網絡在2104年提出後，越來越多的關於圖像分割的深度學習網絡被提出，相比傳統方法，這些網絡效果更好，運算速度更快，已經能成熟的運用在自然圖像上。語義分割顯然已經是計算機視覺領域的一個熱門研究領域，也是通往實現完全場景理解的道路之一，被廣泛應用於無人駕駛、人機交互、醫療圖像、計算攝影、圖像搜索引擎、增強現實等應用領域。語義分割是像素級分類問題，將同一類物體像素點歸爲一類，如圖所示。

左：輸入圖像，右：輸出分割圖像

          存在的挑戰：1.池化或者卷積步長造成的特徵圖分辨率減小；2.圖像中存在不同尺度的目標；3.錯誤匹配關係；4.類別混淆；5.類別不明顯。

          方法：1.dilated convolution；2.圖像金字塔；3.編碼解碼結構；4.級聯結構；5.空間金字塔池化。

1.數據集

2D數據集

1.1 PASCAL Visual Object Classes (VOC)

1.2 PASCAL Context

1.3 PASCAL Part

1.4 Semantic Boundaries Dataset (SBD)

1.5 Microsoft Common Objects in Context (COCO)

1.6 SYNTHetic Collection of Imagery and Annotations (SYNTHIA)

1.7 Cityscapes

1.8 CamVid 

1.9 Youtube-Objects

1.10 Adobe’s Portrait Segmentation Adobe’s Portrait Segmentation

1.11 Materials in Context (MINC)

1.12 Densely-Annotated VIdeo Segmentation (DAVIS)

1.13 Stanford background

1.14 SiftFlow

2.5D數據集

1.15 NYUDv2

1.16 SUN3D

1.17 SUNRGBD

1.18 RGB-D Object Dataset

3D數據集

1.19 ShapeNet Part

1.20 Stanford 2D-3D-S

1.21 A Benchmark for 3D Mesh Segmentation

1.22 Sydney Urban Objects Dataset

1.23 Large-Scale Point Cloud Classification Benchmark

2.圖像標註工具

2.1 labelme: Image Annotation Tool with Python

2.2 labelImgPlus

2.3 PS

2.4 OpenSurfaces Segmentation UI

2.5 ImageSegmentation

2.6 JS Segment Annotator

3. Papers

2017

LinkNet

https://arxiv.org/pdf/1707.03718.pdf

https://github.com/e-lab/LinkNet

ICNet

https://arxiv.org/pdf/1704.08545.pdf

https://github.com/hszhao/ICNet

DeepLabv3

https://arxiv.org/pdf/1706.05587v3.pdf

Mask-RCNN

https://arxiv.org/pdf/1703.06870.pdf

https://github.com/jasjeetIM/Mask-RCNN

ERFNet

http://www.robesafe.uah.es/personal/eduardo.romera/pdfs/Romera17iv.pdf

https://github.com/Eromera/erfnet

Large Kernel Matters

https://arxiv.org/pdf/1703.02719

2016

Fully-Convolutional Network (FCN)

https://arxiv.org/pdf/1605.06211.pdf

https://github.com/shelhamer/fcn.berkeleyvision.org

DeepLab

https://arxiv.org/pdf/1606.00915.pdf

https://bitbucket.org/deeplab/deeplab-public/

ENet

https://arxiv.org/pdf/1606.02147.pdf

https://github.com/TimoSaemann/ENet

PixelNet

https://arxiv.org/pdf/1609.06694.pdf

https://github.com/aayushbansal/PixelNet

RefineNet

https://arxiv.org/pdf/1611.06612.pdf

https://github.com/guosheng/refinenet

PSPNet

https://arxiv.org/pdf/1612.01105.pdf

https://github.com/hszhao/PSPNet

FCIS

https://arxiv.org/pdf/1611.07709.pdf

https://github.com/msracver/FCIS

MultiNet

https://arxiv.org/pdf/1612.07695.pdf

https://github.com/MarvinTeichmann/MultiNet

2015

U-Net

https://arxiv.org/pdf/1505.04597.pdf

https://lmb.informatik.uni-freiburg.de/people/ronneber/u-net/

SegNet

https://arxiv.org/pdf/1511.00561.pdf

https://github.com/alexgkendall/caffe-segnet

DilatedNet

https://arxiv.org/pdf/1511.07122.pdf

https://github.com/nicolov/segmentation_keras

DeepMask

https://arxiv.org/pdf/1506.06204.pdf

https://github.com/facebookresearch/deepmask

CRFasRNN

http://www.robots.ox.ac.uk/%7Eszheng/papers/CRFasRNN.pdf

https://github.com/torrvision/crfasrnn

Dilated convolution

https://arxiv.org/pdf/1511.07122.pdf

https://github.com/fyu/dilation

DeconvNet

https://arxiv.org/pdf/1505.04366.pdf

https://github.com/HyeonwooNoh/DeconvNet

MNC

https://arxiv.org/pdf/1512.04412.pdf

https://github.com/daijifeng001/MNC

Zoomout Semantic Segmentation

https://www.cv-foundation.org/openaccess/content_cvpr_2015/papers/Mostajabi_Feedforward_Semantic_Segmentation_2015_CVPR_paper.pdf

https://bitbucket.org/m_mostajabi/zoom-out-release

4.Blog

A 2017 Guide to Semantic Segmentation with Deep Learning

Semantic Segmentation using Fully Convolutional Networks over the years