CVPR2022论文速递（2022.4.26）！共16篇！Transformer+分割+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
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 : 26 Apr 2022
total number : 16
语义分割/Segmentation - 3 篇
Multi-Head Distillation for Continual Unsupervised Domain Adaptation in Semantic Segmentation
标题：语义分割中的持续无监督域适应的多头蒸馏
论文/Paper: http://arxiv.org/pdf/2204.11667
代码/Code: None
Unsupervised Hierarchical Semantic Segmentation with Multiview Cosegmentation and Clustering Transformers
标题：无监督的分层语义分割，具有多视图coseposepation和聚类Transformers
论文/Paper: http://arxiv.org/pdf/2204.11432
代码/Code: https://github.com/twke18/HSG
Transfer Learning from Synthetic In-vitro Soybean Pods Dataset for In-situ Segmentation of On-branch Soybean Pod
标题：从合成的体外大豆豆荚数据集中迁移学习数据集，以用于分支的原位分割大豆豆荚
论文/Paper: http://arxiv.org/pdf/2204.10902
代码/Code: None
超分/Super-Resolution - 1 篇
IMDeception: Grouped Information Distilling Super-Resolution Network
标题：IMDeception：分组的信息蒸馏超分辨率网络
论文/Paper: http://arxiv.org/pdf/2204.11463
代码/Code: None
Transformers - - 2 篇
Unsupervised Hierarchical Semantic Segmentation with Multiview Cosegmentation and Clustering Transformers
标题：无监督的分层语义分割，具有多视图coseposepation和聚类Transformers
论文/Paper: http://arxiv.org/pdf/2204.11432
代码/Code: https://github.com/twke18/HSG
VISTA: Vision Transformer enhanced by U-Net and Image Colorfulness Frame Filtration for Automatic Retail Checkout
标题：Vista：视觉Transformers通过U-NET和图像彩色框架过滤增强了自动零售结帐
论文/Paper: http://arxiv.org/pdf/2204.11024
代码/Code: None
轨迹预测/Trajectory Prediction - 1 篇
Goal-driven Self-Attentive Recurrent Networks for Trajectory Prediction
标题：轨迹预测的目标驱动的Self-Attentive的反复网络
论文/Paper: http://arxiv.org/pdf/2204.11561
代码/Code: None
人脸相关 / Face - 1 篇
EMOCA: Emotion Driven Monocular Face Capture and Animation
标题：EMOCA：情绪驱动的单目人脸捕捉和动画
论文/Paper: http://arxiv.org/pdf/2204.11312
代码/Code: None
点云/Point Clouds - 1 篇
Surface Reconstruction from Point Clouds by Learning Predictive Context Priors
标题：通过学习预测上下文先验，从点云进行重建
论文/Paper: http://arxiv.org/pdf/2204.11015
代码/Code: None
其他/Other - 8 篇
Proto2Proto: Can you recognize the car, the way I do?
标题：Proto2Proto：您能以我的方式认出汽车吗？
论文/Paper: http://arxiv.org/pdf/2204.11830
代码/Code: None
Multi-Layer Modeling of Dense Vegetation from Aerial LiDAR Scans
标题：空中激光扫描的茂密植被的多层建模
论文/Paper: http://arxiv.org/pdf/2204.11620
代码/Code: https://github.com/ekalinicheva/multi_layer_vegetation.
Surpassing the Human Accuracy: Detecting Gallbladder Cancer from USG Images with Curriculum Learning
标题：超越人类的准确性：从USG图像中检测胆囊癌
论文/Paper: http://arxiv.org/pdf/2204.11433
代码/Code: None
Can domain adaptation make object recognition work for everyone?
标题：域适应能否使对象识别对每个人都有效吗？
论文/Paper: http://arxiv.org/pdf/2204.11122
代码/Code: None
Investigating Neural Architectures by Synthetic Dataset Design
标题：通过合成数据集设计调查神经体系结构
论文/Paper: http://arxiv.org/pdf/2204.11045
代码/Code: None
Revealing Occlusions with 4D Neural Fields
标题：揭示4D神经场的闭塞
论文/Paper: http://arxiv.org/pdf/2204.10916
代码/Code: None
Identity Preserving Loss for Learned Image Compression
标题：为学习图像压缩的身份保留损失
论文/Paper: http://arxiv.org/pdf/2204.10869
代码/Code: None
Towards Data-Free Model Stealing in a Hard Label Setting
标题：朝着硬标签设置中的无数据模型窃取
论文/Paper: http://arxiv.org/pdf/2204.11022
代码/Code: None