bytehuang0912 / awesome-real-time-semantic-segmentation

Awesome-Real-time-Semantic-Segmentation

实时语义分割(Real-time Semantic Segmentation)

• 实时语义分割方法(Real-time Semantic Segmentation Methods)
  • 单分支网络(Single-branch Network)
  • 多分支网络(Two-branch Network)
  • 单分支网络(Multi-branch Network)
  • U型网络(U-shape Network)
  • U型网络(U-shape Network)
• 数据集(Dataset)
• 测速代码(Latency Measurement Code)

实时语义分割方法(Real-time Semantic Segmentation Methods)

单分支网络(Single-branch Network)

方法	标题	论文	代码	发表期刊或会议	基础框架	应用场景	发表年份	数据集
ENet	ENet: A Deep Neural Network Architecture for Real-Time Semantic Segmentation	Paper	Code	Arxiv	CNN	Desk GPU	2016	Cityscapes, CamVid, SUN
DABNet	DABNet: Depth-wise Asymmetric Bottleneck for Real-time Semantic Segmentation	Paper	Code	BMVC	CNN	Desk GPU	2019	Cityscapes, CamVid
SegFormer	SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers	Paper	Code	NeurIPS	Transformer	Desk GPU	2021	Cityscapes, ADE20K, PascalContext, PascalVOC, COCO-Stuff, iSAID
SegNext	SegNeXt: Rethinking Convolutional Attention Design for Semantic Segmentation	Paper	Code	NeurIPS	CNN	Desk GPU	2022	Cityscapes, ADE20K, PascalContext, PascalVOC, COCO-Stuff, iSAID
AFFormer	Head-Free Lightweight Semantic Segmentation with Linear Transformer	Paper	Code	AAAI	Transformer	Desk GPU	2023	Cityscapes, ADE20K, COCO-Stuff

双分支网络(Two-branch Network)

方法	标题	论文	代码	发表期刊或会议	基础框架	应用场景	发表年份	数据集
BiSeNet	BiSeNet: Bilateral Segmentation Network for Real-time Semantic Segmentation	Paper	Code	ECCV	CNN	Desk GPU	2018	Cityscapes, CamVid, COCO-Stuff
Fast-SCNN	Fast-SCNN: Fast Semantic Segmentation Network	Paper	Code	BMVC	CNN	Desk GPU	2019	Cityscapes
BiSeNetV2	BiSeNet V2: Bilateral Network with Guided Aggregation for Real-time Semantic Segmentation	Paper	Code	IJCV	CNN	Desk GPU	2021	Cityscapes, CamVid, COCO-Stuff
STDC	Rethinking BiSeNet For Real-time Semantic Segmentation	Paper	Code	CVPR	CNN	Desk GPU	2021	Cityscapes, CamVid
DDRNet	Deep Dual-resolution Networks for Real-time and Accurate Semantic Segmentation of Traffic Scenes	Paper	Code	TIP	CNN	Desk GPU	2022	Cityscapes, CamVid, COCO-Stuff
RTFormer	Rethinking BiSeNet For Real-time Semantic Segmentation	Paper	Code	NeurIPS	CNN	Desk GPU	2022	Cityscapes, CamVid, ADE20K, COCO-Stuff
SeaFormer	SEAFORMER: SQUEEZE-ENHANCED AXIAL TRANSFORMER FOR MOBILE SEMANTIC SEGMENTATION	Paper	Code	ICLR	Hybrid	Mobile CPU	2023	Cityscapes, ADE20K, PascalContext, COCO-Stuff

多分支网络(Multi Branch Network)

方法	标题	论文	代码	发表期刊或会议	基础框架	应用场景	发表年份	数据集
ICNet	ICNet for Real-Time Semantic Segmentation on High-Resolution Images	Paper	Code	ECCV	CNN	Desk GPU	2018	Cityscapes, CamVid, COCO-Stuff
ESPNet	ESPNet: Efficient Spatial Pyramid of Dilated Convolutions for Semantic Segmentation	Paper	Code	ECCV	CNN	Desk GPU	2018	Cityscapes, PascalVOC, Mapillary
DFANet	DFANet: Deep Feature Aggregation for Real-Time Semantic Segmentation	Paper		CVPR	CNN	Desk GPU	2019	Cityscapes, CamVid
PIDNet	PIDNet: A Real-time Semantic Segmentation Network Inspired by PID Controllers	Paper	Code	CVPR	CNN	Desk GPU	2023	Cityscapes, CamVid,PASCAL Context

U型网络(U-shape Network)

方法	标题	论文	代码	发表期刊或会议	基础框架	应用场景	发表年份	数据集
SwiftNet	In Defense of Pre-trained ImageNet Architectures for Real-time Semantic Segmentation of Road-driving Images	Paper	Code	CVPR	CNN	Desk GPU	2019	Cityscapes, CamVid
ShelfNet	ShelfNet for Fast Semantic Segmentation	Paper	Code	ICCV(Workshop)	CNN	Desk GPU	2019	Cityscapes, PascalContext, PascalVOC
SFNet	Semantic Flow for Fast and Accurate Scene Parsing	Paper	Code	ECCV	CNN	Desk GPU	2020	Cityscapes, CamVid, ADE20K, PascalContext
HyperSeg	HyperSeg: Patch-wise Hypernetwork for Real-time Semantic Segmentation	Paper	Code	CVPR	CNN	Desk GPU	2021	Cityscapes, CamVid, PascalVOC
TopFormer	TopFormer: Token Pyramid Transformer for Mobile Semantic Segmentation	Paper	Code	CVPR	Hybrid	Mobile CPU	2022	Cityscapes, ADE20K, PascalContext, COCO-Stuff

NAS网络(NAS Network)

方法	标题	论文	代码	发表期刊或会议	基础框架	应用场景	发表年份	数据集
DF-Seg	Partial Order Pruning: for Best Speed/Accuracy Trade-off in Neural Architecture Search	Paper		CVPR	CNN	Desk GPU	2019	Cityscapes
FasterSeg	FASTERSEG: SEARCHING FOR FASTER REAL-TIME SEMANTIC SEGMENTATION	Paper	Code	ICLR	CNN	Desk GPU	2020	Cityscapes, CamVid, BDD
RT-Seg	Towards Real-Time Segmentation on the Edge	Paper		AAAI	CNN	Mobile GPU	2023	Cityscapes, ADE20K, PascalVOC
Pruning Parameterization	Pruning Parameterization with Bi-level Optimization for Efficient Semantic Segmentation on the Edge	Paper		CVPR	CNN	Mobile GPU	2023	Cityscapes, ADE20K, PascalVOC

数据集(Dataset)

数据集	链接	年份	类别数	分辨率	图像数量	类型
Cityscapes	Link	2012	19	2048x1024	5000	自动驾驶
CamVid	Link	2009	11	960x720	701	自动驾驶
ADE20K	Link	2017	150	不固定	22210	通用场景
COCOStuff-10K	Link	2018	171	不固定	10000	通用场景
PASCAL VOC 2012	Link	2012	20	不固定	13487	通用场景
PascalContext	Link	2015	59	不固定	10103	通用场景

测速代码(Latency Measurement Code)

Link