CVPR2022论文速递（2022.5.12）！共6篇！

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 May 2022

total number : 5

分类 / Classification - 1 篇

Few-Shot Image Classification Benchmarks are Too Far From Reality: Build Back Better with Semantic Task Sampling

标题：少样本图像分类基准离现实太远：通过语义任务采样更好地恢复原状

• 论文/Paper: http://arxiv.org/pdf/2205.05155
• 代码/Code: None

语义分割/Segmentation - 1 篇

Scene Consistency Representation Learning for Video Scene Segmentation

标题：视频场景细分的场景一致性表示学习

• 论文/Paper: http://arxiv.org/pdf/2205.05487
• 代码/Code: https://github.com/TencentYoutuResearch/SceneSegmentation-SCRL

其他/Other - 3 篇

Revisiting Random Channel Pruning for Neural Network Compression

标题：重新考虑神经网络压缩的随机通道修剪

• 论文/Paper: http://arxiv.org/pdf/2205.05676
• 代码/Code: https://github.com/ofsoundof/random_channel_pruning

Video-ReTime: Learning Temporally Varying Speediness for Time Remapping

标题：Video-ReTime：学习时间变化的时间变化

• 论文/Paper: http://arxiv.org/pdf/2205.05609
• 代码/Code: None

TDT: Teaching Detectors to Track without Fully Annotated Videos

标题：TDT：教探测器无需完全注释的视频而进行跟踪

• 论文/Paper: http://arxiv.org/pdf/2205.05583
• 代码/Code: None