CVPR2022论文速递（2022.4.12）！共24篇！GAN/transformer/超分等

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

Updated on : 12 Apr 2022

total number : 24

分类 / Classification - 1 篇

Joint Distribution Matters: Deep Brownian Distance Covariance for Few-Shot Classification

标题：联合分配事项：少量分类的深褐色距离协方差

• 论文/Paper: http://arxiv.org/pdf/2204.04567
• 代码/Code: None

语义分割/Segmentation - 1 篇

Video K-Net: A Simple, Strong, and Unified Baseline for Video Segmentation

标题：视频K-Net：视频分割的简单，强大和统一的基线

• 论文/Paper: http://arxiv.org/pdf/2204.04656
• 代码/Code: https://github.com/lxtGH/Video-K-Net

GAN - 2 篇

Commonality in Natural Images Rescues GANs: Pretraining GANs with Generic and Privacy-free Synthetic Data

标题：自然图像中的共性救援GANS：预先利用通用和无自由的合成数据

• 论文/Paper: http://arxiv.org/pdf/2204.04950
• 代码/Code: None

Commonality in Natural Images Rescues GANs: Pretraining GANs with Generic and Privacy-free Synthetic Data

标题：自然图像中的共性救援GANS：预先利用通用和无自由的合成数据

• 论文/Paper: http://arxiv.org/pdf/2204.04950
• 代码/Code: None

超分/Super-Resolution - 1 篇

Learning Trajectory-Aware Transformer for Video Super-Resolution

标题：用于视频超分辨率的学习轨迹感知Transformer

• 论文/Paper: http://arxiv.org/pdf/2204.04216
• 代码/Code: https://github.com/researchmm/TTVSR

Transformers - - 4 篇

Consistency Learning via Decoding Path Augmentation for Transformers in Human Object Interaction Detection

标题：通过解码路径增强用于人类对象交互检测的Transformer的一致性学习

• 论文/Paper: http://arxiv.org/pdf/2204.04836
• 代码/Code: https://github.com/mlvlab/CPChoi.

Multimodal Transformer for Nursing Activity Recognition

标题：用于护理活动识别的多峰Transformer

• 论文/Paper: http://arxiv.org/pdf/2204.04564
• 代码/Code: \url{https://github.com/Momilijaz96/MMT_for_NCRC}.

Multimodal Transformer for Nursing Activity Recognition

标题：用于护理活动识别的多峰Transformer

• 论文/Paper: http://arxiv.org/pdf/2204.04564
• 代码/Code: \url{https://github.com/Momilijaz96/MMT_for_NCRC}.

Learning Trajectory-Aware Transformer for Video Super-Resolution

标题：用于视频超分辨率的学习轨迹感知Transformer

• 论文/Paper: http://arxiv.org/pdf/2204.04216
• 代码/Code: https://github.com/researchmm/TTVSR

多模态 / Multimodal - 3 篇

XMP-Font: Self-Supervised Cross-Modality Pre-training for Few-Shot Font Generation

标题：XMP-FONT：少量字体生成的自我监督的跨模型预训练

• 论文/Paper: http://arxiv.org/pdf/2204.05084
• 代码/Code: None

Robust Cross-Modal Representation Learning with Progressive Self-Distillation

标题：具有逐步自蒸馏的强大跨莫代代表学习

• 论文/Paper: http://arxiv.org/pdf/2204.04588
• 代码/Code: None

Multimodal Transformer for Nursing Activity Recognition

标题：用于护理活动识别的多峰Transformer

• 论文/Paper: http://arxiv.org/pdf/2204.04564
• 代码/Code: \url{https://github.com/Momilijaz96/MMT_for_NCRC}.

姿态估计/Pose Estimation - 1 篇

Focal Length and Object Pose Estimation via Render and Compare

标题：通过渲染和比较焦距和对象姿态估计

• 论文/Paper: http://arxiv.org/pdf/2204.05145
• 代码/Code: http://github.com/ponimatkin/focalpose

检索/Image Retrieval - 1 篇

Beyond Cross-view Image Retrieval: Highly Accurate Vehicle Localization Using Satellite Image

标题：超越巧克力视图检索：使用卫星图像的高精度车辆本地化

• 论文/Paper: http://arxiv.org/pdf/2204.04752
• 代码/Code: None

NeRF - 1 篇

NAN: Noise-Aware NeRFs for Burst-Denoising

标题：NaN：噪音感知的NERF用于爆发去噪

• 论文/Paper: http://arxiv.org/pdf/2204.04668
• 代码/Code: None

深度估计/Depth Estimation - 1 篇

HiMODE: A Hybrid Monocular Omnidirectional Depth Estimation Model

标题：HIMODE：混合单眼全向深度估计模型

• 论文/Paper: http://arxiv.org/pdf/2204.05007
• 代码/Code: None

其他/Other - 12 篇

Single-Photon Structured Light

标题：单光子结构光

• 论文/Paper: http://arxiv.org/pdf/2204.05300
• 代码/Code: None

Pyramid Grafting Network for One-Stage High Resolution Saliency Detection

标题：金字塔嫁接网络用于一级高分辨率显着性检测

• 论文/Paper: http://arxiv.org/pdf/2204.05041
• 代码/Code: None

Structure-Aware Motion Transfer with Deformable Anchor Model

标题：具有可变形锚模型的结构感知运动传输

• 论文/Paper: http://arxiv.org/pdf/2204.05018
• 代码/Code: None

SOS! Self-supervised Learning Over Sets Of Handled Objects In Egocentric Action Recognition

标题：SOS！自我监督学习在Egentric行动识别中的处理对象集

• 论文/Paper: http://arxiv.org/pdf/2204.04796
• 代码/Code: None

Reasoning with Multi-Structure Commonsense Knowledge in Visual Dialog

标题：在视觉对话中的多结构致辞知识推理

• 论文/Paper: http://arxiv.org/pdf/2204.04680
• 代码/Code: None

Learning Pixel-Level Distinctions for Video Highlight Detection

标题：学习像素级别的视频突出显示检测

• 论文/Paper: http://arxiv.org/pdf/2204.04615
• 代码/Code: None

Explaining Deep Convolutional Neural Networks via Latent Visual-Semantic Filter Attention

标题：通过潜在视觉语义滤波器注意解释深卷积神经网络

• 论文/Paper: http://arxiv.org/pdf/2204.04601
• 代码/Code: None

DeepLIIF: An Online Platform for Quantification of Clinical Pathology Slides

标题：DEEPLIIF：用于量化临床病理学幻灯片的在线平台

• 论文/Paper: http://arxiv.org/pdf/2204.04494
• 代码/Code: None

ManiTrans: Entity-Level Text-Guided Image Manipulation via Token-wise Semantic Alignment and Generation

标题：曼特拉人：通过令牌语义对齐和生成实体级文本引导图像操纵

• 论文/Paper: http://arxiv.org/pdf/2204.04428
• 代码/Code: None

FedCorr: Multi-Stage Federated Learning for Label Noise Correction

标题：FEDCORR：用于标签噪声校正的多级联合学习

• 论文/Paper: http://arxiv.org/pdf/2204.04677
• 代码/Code: https://github.com/Xu-Jingyi/FedCorr

Adaptive Differential Filters for Fast and Communication-Efficient Federated Learning

标题：自适应差分滤波器，用于快速和通信高效的联合学习

• 论文/Paper: http://arxiv.org/pdf/2204.04424
• 代码/Code: None

The Two Dimensions of Worst-case Training and the Integrated Effect for Out-of-domain Generalization

标题：最坏情况训练的两个维度和域外概括的综合效果

• 论文/Paper: http://arxiv.org/pdf/2204.04384
• 代码/Code: None