CVPR2022论文速递（2022.4.11）！共12篇！跟踪/transformer/对比学习等

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：

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 : 11 Apr 2022

total number : 12

目标跟踪/Object Tracking - 1 篇

Visible-Thermal UAV Tracking: A Large-Scale Benchmark and New Baseline

标题：可见热无人机跟踪：大型基准和新基线

• 论文/Paper: http://arxiv.org/abs/2204.04120

• 代码/Code: None

Transformers - - 1 篇

Gravitationally Lensed Black Hole Emission Tomography

标题：引力镜头的黑洞发射断层扫描

• 论文/Paper: http://arxiv.org/abs/2204.03715

• 代码/Code: None

对比学习/Contrastive Learning - 1 篇

Probabilistic Representations for Video Contrastive Learning

标题：视频对比学习的概率表

• 论文/Paper: http://arxiv.org/abs/2204.03946

• 代码/Code: None

其他/Other - 9 篇

Dancing under the stars: video denoising in starlight

标题：在星空下跳舞：在星光中的视频去噪

• 论文/Paper: http://arxiv.org/abs/2204.04210

• 代码/Code: None

General Incremental Learning with Domain-aware Categorical Representations

标题：域感知分类表示的一般增量学习

• 论文/Paper: http://arxiv.org/abs/2204.04078

• 代码/Code: None

Identifying Ambiguous Similarity Conditions via Semantic Matching

标题：通过语义匹配识别模糊的相似条件

• 论文/Paper: http://arxiv.org/abs/2204.04053

• 代码/Code: None

Does Robustness on ImageNet Transfer to Downstream Tasks?

标题：在想象网转移到下游任务的鲁棒性吗？

• 论文/Paper: http://arxiv.org/abs/2204.03934

• 代码/Code: None

Deep Hyperspectral-Depth Reconstruction Using Single Color-Dot Projection

标题：使用单色点投影深度高光谱深度重建

• 论文/Paper: http://arxiv.org/abs/2204.03929

• 代码/Code: None

CD$^2$-pFed: Cyclic Distillation-guided Channel Decoupling for Model Personalization in Federated Learning

标题：CD $ ^ 2 $ -pfed：循环蒸馏引导的通道解耦，用于联邦学习中的模型个性化

• 论文/Paper: http://arxiv.org/abs/2204.03880

• 代码/Code: None

Reusing the Task-specific Classifier as a Discriminator: Discriminator-free Adversarial Domain Adaptation

标题：将特定于任务特定的分类器重新定义为判别：无鉴别者的对抗域适应

• 论文/Paper: http://arxiv.org/abs/2204.03838

• 代码/Code: https://github.com/xiaoachen98/DALN

TorMentor: Deterministic dynamic-path, data augmentations with fractals

标题：Tormentor：确定性动态路径，带分形的数据增强

• 论文/Paper: http://arxiv.org/abs/2204.03776

• 代码/Code: None

TemporalUV: Capturing Loose Clothing with Temporally Coherent UV Coordinates

标题：Temporaluv：用时间相干的UV坐标捕获松散的衣服

• 论文/Paper: http://arxiv.org/abs/2204.03671

• 代码/Code: None