CVPR2022论文速递（2022.4.20）！共13篇！Transformer+Person ReID

整理：AI算法与图像处理

CVPR2022论文和代码整理：https://github.com/DWCTOD/CVPR2022-Papers-with-Code-Demo

欢迎关注公众号 AI算法与图像处理，获取更多干货：

大家好, 最近正在优化每周分享的CVPR论文, 目前考虑按照不同类别去分类,方便不同方向的小伙伴挑选自己感兴趣的论文哈

欢迎大家留言其他想法, 合适的话会采纳哈! 求个三连支持一波哈

Updated on : 21 Apr 2022

total number : 13

分类 / Classification - 1 篇

Self-supervised Learning for Sonar Image Classification

标题：声纳图像分类的自我监督学习

论文/Paper: http://arxiv.org/pdf/2204.09323
代码/Code: https://github.com/agrija9/ssl-sonar-images

目标跟踪/Object Tracking - 1 篇

Multi-Camera Multiple 3D Object Tracking on the Move for Autonomous Vehicles

标题：自动车辆移动的多摄像机多维对象跟踪

论文/Paper: http://arxiv.org/pdf/2204.09151
代码/Code: None

语义分割/Segmentation - 1 篇

Unsupervised Domain Adaptation for Cardiac Segmentation: Towards Structure Mutual Information Maximization

标题：心脏分割的无监督域适应：朝着结构相互信息最大化

论文/Paper: http://arxiv.org/pdf/2204.09334
代码/Code: https://github.com/LOUEY233/Toward-Mutual-Information}{https://github.com/LOUEY233/Toward-Mutual-Information

Transformers - - 2 篇

NFormer: Robust Person Re-identification with Neighbor Transformer

标题：NFormer：用Neighbor Transformer进行鲁棒的行人重识别

论文/Paper: http://arxiv.org/pdf/2204.09331
代码/Code: \url{https://github.com/haochenheheda/NFormer}.

Human-Object Interaction Detection via Disentangled Transformer

标题：通过解耦Transformer人-机交互检测

论文/Paper: http://arxiv.org/pdf/2204.09290
代码/Code: None

行人重识别/Person Re-Identification - 1 篇

NFormer: Robust Person Re-identification with Neighbor Transformer

标题：标题：NFormer：用Neighbor Transformer进行鲁棒的行人重识别

论文/Paper: http://arxiv.org/pdf/2204.09331
代码/Code: \url{https://github.com/haochenheheda/NFormer}.

其他/Other - 8 篇

GazeOnce: Real-Time Multi-Person Gaze Estimation

标题：GazeOnce：实时多人瞳孔估计

论文/Paper: http://arxiv.org/pdf/2204.09480
代码/Code: None

Hephaestus: A large scale multitask dataset towards InSAR understanding

标题：Hephaestus：一个大规模的多任务数据集，旨在理解

论文/Paper: http://arxiv.org/pdf/2204.09435
代码/Code: None

Epistemic Uncertainty-Weighted Loss for Visual Bias Mitigation

标题：视觉偏差缓解的认知不确定性加权损失

论文/Paper: http://arxiv.org/pdf/2204.09389
代码/Code: None

Reinforced Structured State-Evolution for Vision-Language Navigation

标题：用于视觉语言导航的增强结构状态演变

论文/Paper: http://arxiv.org/pdf/2204.09280
代码/Code: None

SpiderNet: Hybrid Differentiable-Evolutionary Architecture Search via Train-Free Metrics

标题：Spidernet：混合可分辨率 - 进化架构通过渔酒指标进行搜索

论文/Paper: http://arxiv.org/pdf/2204.09320
代码/Code: None

A Deeper Look into Aleatoric and Epistemic Uncertainty Disentanglement

标题：更深层次的观察炼热和认知不确定性解剖学

论文/Paper: http://arxiv.org/pdf/2204.09308
代码/Code: None

Does Interference Exist When Training a Once-For-All Network?

标题：再训练一次所有网络时都存在干扰吗？

论文/Paper: http://arxiv.org/pdf/2204.09210
代码/Code: https://github.com/Jordan-HS/RSS-Interference-CVPRW2022.

Importance is in your attention: agent importance prediction for autonomous driving

标题：重要性在您的注意力：Agent Instance预测自动驾驶

论文/Paper: http://arxiv.org/pdf/2204.09121
代码/Code: None