TransRAC: Encoding Multi-scale Temporal Correlation with Transformers for Repetitive Action Counting，CVPR 2022 Oral

论文：https://arxiv.org/abs/2204.01018 [2204.01018] TransRAC: Encoding Multi-scale Temporal Correlation with Transformers for Repetitive Action Counting (arxiv.org)

代码：https://github.com/SvipRepetitionCounting/TransRAC GitHub - SvipRepetitionCounting/TransRAC: (CVPR 2022 Oral) Official implemention: TransRAC数据集：https://svip-lab.github.io/dataset/RepCount_dataset.html

Shanghaitech Vision and Intelligent Perception(SVIP) LAB (svip-lab.github.io)

Learning Part Segmentation through Unsupervised Domain Adaptation from Synthetic Vehicles，CVPR 2022 Oral

论文：https://arxiv.org/abs/2103.14098

代码：GitHub - qliu24/render-3d-segmentation

数据：UDA-Part: A Part Segmentation Dataset Based on 3D Computer Graphics Models | UDA-Part

Semantic-Aware Domain Generalized Segmentation，CVPR 2022 Oral

论文：[2204.00822] Semantic-Aware Domain Generalized Segmentation (arxiv.org)
代码：

GitHub - leolyj/SAN-SAW: This is the code related to "Semantic-Aware Domain Generalized Segmentation" (CVPR 2022)

MAXIM: Multi-Axis MLP for Image Processing，CVPR 2022 Oral

论文：[2201.02973] MAXIM: Multi-Axis MLP for Image Processing (arxiv.org)
代码：https://github.com/google-research/maxim

Correlation Verification for Image Retrieval，CVPR 2022 Oral

论文：[2204.01458] Correlation Verification for Image Retrieval (arxiv.org)
代码：GitHub - sungonce/CVNet: Official PyTorch Implementation of Correlation Verifcation for Image Retrieval, CVPR 2022 (Oral Presentation)

Rethinking Semantic Segmentation: A Prototype View，CVPR 2022 Oral

论文：[2203.15102] Rethinking Semantic Segmentation: A Prototype View (arxiv.org)
代码：GitHub - tfzhou/ProtoSeg: CVPR2022 (Oral) - Rethinking Semantic Segmentation: A Prototype View

Dual-AI: Dual-path Action Interaction Learning for Group Activity Recognition，CVPR 2022 Oral

论文：[2204.02148] Dual-AI: Dual-path Actor Interaction Learning for Group Activity Recognition (arxiv.org)
代码：Dual-AI: Dual-path Actor Interaction Learning for Group Activity Recognition (mingfei.info)

GAN-Supervised Dense Visual Alignment，CVPR 2022 Oral

论文：[2112.05143] GAN-Supervised Dense Visual Alignment (arxiv.org)
代码：https://github.com/wpeebles/gangealing
项目：GAN-Supervised Dense Visual Alignment (wpeebles.com)

Multi-View Depth Estimation by Fusing Single-View Depth Probability with Multi-View Geometry，CVPR 2022 Oral

论文：[2112.08177] Multi-View Depth Estimation by Fusing Single-View Depth Probability with Multi-View Geometry (arxiv.org)
代码：GitHub - baegwangbin/MaGNet: (CVPR 2022 - oral) Multi-View Depth Estimation by Fusing Single-View Depth Probability with Multi-View Geometry

SeqFormer：Sequential Transformer for Video Instance Segmentation, ECCV 2022 Oral

SeqFormer：https://arxiv.org/abs/2112.08275

IDOL：https://arxiv.org/abs/2207.10661

官方代码地址：https://github.com/wjf5203/VNex

ECCV 2022 Oral | 满分论文！视频实例分割新SOTA：SeqFormer & IDOL - 知乎 (zhihu.com)

In Defense of Online Models for Video Instance Segmentation, ECCV 2022 Oral

In Defense of Online Models for Video Instance Segmentation - 知乎 (zhihu.com)

https://arxiv.org/abs/2207.10661

Large-scale Unsupervised Semantic Segmentation，TPAMI2022

数据集下载：github.com/LUSSeg/ImageNet-S

主流方法代码：github.com/LUSSeg/ImageNetSegModel

论文所提出的方法：github.com/LUSSeg/PASS

论文地址：arxiv.org/pdf/2106.03149.pdf

TPAMI2022: 大规模无监督语义分割（LUSS）及其数据集 Image-S

Discovering Object Masks with Transformers for Unsupervised Semantic Segmentation

论文地址：https://arxiv.org/pdf/2206.06363.pdf

开源代码：https://github.com/wvangansbeke/MaskDistill

无监督语义分割-MaskDistill：用Transformer来挖掘先验，不需要任何标注数据达到SOTA

StructToken : Rethinking Semantic Segmentation with Structural Prior

https://arxiv.org/pdf/2203.12612.pdf

语义分割新范式：上海 AI Lab 联合北邮、商汤提出 StructToken

TokenMix: Rethinking Image Mixing for Data Augmentation in Vision Transformers，ECCV 2022

论文地址：https://arxiv.org/abs/2207.08409[1]

代码地址：https://github.com/Sense-X/TokenMix[2]

ECCV 2022 吊打MixUp和CutMix！MMLab&商汤提出超强数据增强策略《TokenMix》！ - 知乎 (zhihu.com)

Shunted Self-Attention via Multi-Scale Token Aggregation

论文地址：https://arxiv.org/pdf/2111.15193.pdf

代码地址：https://github.com/OliverRensu/Shunted-Transformer

CVPR2022 Oral - Shunted Transformer：全新多尺度视觉 Transformer 主干网络

Tree Energy Loss: Towards Sparsely Annotated Semantic Segmentation

论文地址：https://arxiv.org/pdf/2203.10739.pdf

代码地址：https://github.com/megvii-research/TreeEnergyLoss

CVPR2022-Tree Energy Loss：能够扩展弱监督语义分割中稀疏真值标签的新方法

精读Tree Energy Loss: Towards Sparsely Annotated Semantic Segmentation_格里芬阀门工的博客-CSDN博客

Weakly supervised Semantic Segmentation by Pixel-to-Prototype Contrast

论文地址：https://arxiv.org/pdf/2110.07110.pdf

代码地址：未开源

CVPR2022-Pixel-to-Prototype Contrast：将对比学习应用于弱监督语义分割

Class Re-Activation Maps for Weakly-Supervised Semantic Segmentation， CVPR2022

论文地址：https://arxiv.org/pdf/2203.00962.pdf

代码地址：https://github.com/zhaozhengChen/ReCAM

CVPR2022-Class Re-Activation Maps：用于弱监督语义分割的类重新激活图

Cross-Dataset Collaborative Learning for Semantic Segmentation，AAAI 2022

论文地址：https://arxiv.org/pdf/2103.11351.pdf

代码地址：https://github.com/wanglixilinx/CDCL

AAAI2022-Cross-Dataset Collaborative Learning: 语义分割中的跨数据集协同学习

Cross-Dataset Collaborative Learning for Semantic Segmentation - 知乎

Pyraformer: Low-Complexity Pyramidal Attention for Long-range Time Series Modeling and Forecasting，ICLR 2022 oral

论文地址：https://openreview.net/pdf?id=0EXmFzUn5I
github地址：https://github.com/alipay/Pyraf

ICLR 2022 Oral | 蚂蚁集团用于长程时序建模的开源算法Pyraformer简介 - 知乎

Discovering and Explaining the Representation Bottleneck of DNNs，ICLR 2022 oral

ICLR 2022 Oral论文中得分排名前五的高分论文“发现并证明神经网络表征瓶颈”（得分10,8,8,8） - 知乎

Open-Set Recognition: a Good Closed-Set Classifier is All You Need?，ICLR 2022 oral

论文链接：https://arxiv.org/abs/2110.06207
项目链接：https://github.com/sgvaze/osr_closed_set_all_you_need

相得益彰！提升模型闭集分类精度即可提升开集检测的准确率（ICLR 2022 oral）

PICO: CONTRASTIVE LABEL DISAMBIGUATION FOR PARTIAL LABEL LEARNING， ICLR 2022 best paper

下载地址：https://openreview.net/pdf?id=EhYjZy6e1gJ

本文|业内_ICLR 2022最佳论文解读

Non-Transferable Learning: A New Approach for Model Ownership Verification and Applicability Authorization，ICLR 2022 oral

Non-Transferable Learning: A New Approach for Model Ownership Verification and Applicability Authorization

Non-Transferable Learning: A New Approach for Model Ownership Verification and Applicability Authorization | OpenReview

ICLR 2022 Oral: Non-Transferable Learning(反迁移学习) - 知乎

ICLR2022oral site:zhuanlan.zhihu.com

清华、人大获奖，浙大提名，ICLR 2022杰出论文奖出炉 - 知乎

参考博客