Segmentation 관련 글 목차
2000, Jan 01
- 이 글은 딥러닝을 이용한 Segmentation 모델을 다룬 글들의 목록입니다.
- Segmentation 모델들의 목록은 유명한 기본 모델들과 Cityscape Benchmarks에 등록된 논문들입니다.
- 모델의 분류는
mIoU성능 위주의 모델과Realtime으로 처리 가능한 모델 위주로 분류하였습니다.
Segmentation 관련 내용
- segmentation 학습을 위한 one hot label 생성 및 Loss 연산
- mIoU(Mean Intersection over Union)
- Segmentation의 Entropy 표현 방법
- Multi-Scale Context Aggregation by Dilated Convolutions
- ASPP(Atrous Spatial Pyramid Pooling)
- DenseCRF, Efficient Inference in Fully Connected CRFs with Gaussian Edge Potentials
- Learning Deconvolution Network for Semantic Segmentation
- Multi-scale context aggregation by dilated convolutions
- STEGO (Unsupervised Semantic Segmentation by Distilling Feature Correspondences)
Performance Oriented Sementic Segmentation 모델
- DeepLab_v1 (Semantic Image Segmentation with Deep Convolutional Nets and Fully Connected CRFs)
- DeepLab_v2 (DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs)
- DeepLab_v3 (Rethinking Atrous Convolution for Semantic Image Segmentation)
- DeepLab_v3+ (Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation)
- DFN(Discriminative Feature Network)
- The One Hundred Layers Tiramisu: Fully Convolutional DenseNets for Semantic Segmentation
- DenseASPP for Semantic Segmentation in Street Scenes
- PSPNet(Pyramid Scene Parsing Network)
- Hierarchical Multi-Scale Attention for Semantic Segmentation
- SegFormer, Simple and Efficient Design for Semantic Segmentation with Transformers
Realtime Oriented Sementic Segmentation 모델
- RTSeg: Real-time Semantic Segmentation Comparative Study
- Fully Convolutional Networks for Semantic Segmentation
- U-Net
- ESPNet
- SegNet
- CGNet
- LiteSeg, A Novel Lightweight ConvNet for Semantic Segmentation
- FCHarDNet