CVPR2022论文速递（2022.4.6）！共14篇，已归类

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：​

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 : 6 Apr 2022

total number : 14

数据集/Dataset - 1 篇

ObjectFolder 2.0: A Multisensory Object Dataset for Sim2Real Transfer

标题：ObjectFolder 2.0：SIM2REAL TRANSER的多句子对象数据集

• 论文/Paper: http://arxiv.org/pdf/2204.02389
• 代码/Code: None

目标检测 / Object Detection - 1 篇

Overcoming Catastrophic Forgetting in Incremental Object Detection via Elastic Response Distillation

标题：通过弹性响应蒸馏克服灾难性遗忘

• 论文/Paper: http://arxiv.org/pdf/2204.02136
• 代码/Code: None

GAN - 1 篇

Arbitrary-Scale Image Synthesis

标题：任意尺度图像合成

• 论文/Paper: http://arxiv.org/pdf/2204.02273
• 代码/Code: https://github.com/vglsd/ScaleParty

Transformers - - 4 篇

Multi-View Transformer for 3D Visual Grounding

标题：用于3D视觉接地的多视图变压器

• 论文/Paper: http://arxiv.org/pdf/2204.02174
• 代码/Code: None

Dual-AI: Dual-path Action Interaction Learning for Group Activity Recognition

标题：双AI：组活动识别的双路径交互学习

• 论文/Paper: http://arxiv.org/pdf/2204.02148
• 代码/Code: None

Detector-Free Weakly Supervised Group Activity Recognition

标题：探测器无弱监督集团活动识别

• 论文/Paper: http://arxiv.org/pdf/2204.02139
• 代码/Code: None

Text Spotting Transformers

标题：文本拍摄变形金刚

• 论文/Paper: http://arxiv.org/pdf/2204.01918
• 代码/Code: None

深度估计/Depth Estimation - 1 篇

P3Depth: Monocular Depth Estimation with a Piecewise Planarity Prior

标题：p3depth：先前的分段平面单曲深度估计

• 论文/Paper: http://arxiv.org/pdf/2204.02091
• 代码/Code: None

其他/Other - 6 篇

Rethinking Visual Geo-localization for Large-Scale Applications

标题：重新思考大规模应用的视觉地理定位

• 论文/Paper: http://arxiv.org/pdf/2204.02287
• 代码/Code: None

IRON: Inverse Rendering by Optimizing Neural SDFs and Materials from Photometric Images

标题：铁：通过从光度图像优化神经系统和材料来实现逆渲染

• 论文/Paper: http://arxiv.org/pdf/2204.02232
• 代码/Code: None

SNUG: Self-Supervised Neural Dynamic Garments

标题：舒适：自我监督的神经动态服装

• 论文/Paper: http://arxiv.org/pdf/2204.02219
• 代码/Code: None

Leveraging Equivariant Features for Absolute Pose Regression

标题：绝对姿势回归利用等分意识的功能

• 论文/Paper: http://arxiv.org/pdf/2204.02163
• 代码/Code: None

MonoTrack: Shuttle trajectory reconstruction from monocular badminton video

标题：Monotrack：单眼羽毛球视频的梭轨迹重建

• 论文/Paper: http://arxiv.org/pdf/2204.01899
• 代码/Code: None

Revisiting Near/Remote Sensing with Geospatial Attention

标题：在地理空间关注重新审视/遥感

• 论文/Paper: http://arxiv.org/pdf/2204.01807
• 代码/Code: None