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
ECCV2022论文和代码整理:https://github.com/DWCTOD/ECCV2022-Papers-with-Code-Demo
最新成果demo展示:
CVPR2022 | 视频摩尔纹去除
标题:Video Demoireing with Relation-Based Temporal Consistency
主页:https://daipengwa.github.io/VDmoire_ProjectPage/ 论文:https://arxiv.org/pdf/2204.02957.pdf 代码:https://github.com/CVMI-Lab/VideoDemoireing
摘要:
当使用数码相机拍摄屏幕时,出现颜色失真的摩尔纹会严重降低图像和视频质量。考虑到对捕获视频的需求日益增加,我们研究如何去除视频中这种不受欢迎的波纹图案,即视频去波纹。为此,我们引入了第一个具有专用数据收集管道的手持视频 deoireing 数据集,以确保捕获数据的空间和时间对齐。此外,开发了具有隐式特征空间对齐和选择性特征聚合的基线视频去噪模型,以利用来自附近帧的互补信息来改进帧级视频去噪。更重要的是,我们提出了一种基于关系的时间一致性损失,以鼓励模型直接从真实参考视频中学习时间一致性先验,这有助于产生时间一致的预测并有效地保持帧级质量。大量的实验证明了我们模型的优越性。
最新论文整理
ECCV2022
Updated on : 29 Aug 2022
total number : 6
CMD: Self-supervised 3D Action Representation Learning with Cross-modal Mutual Distillation
- 论文/Paper: http://arxiv.org/pdf/2208.12448
- 代码/Code: https://github.com/maoyunyao/cmd
Bokeh-Loss GAN: Multi-Stage Adversarial Training for Realistic Edge-Aware Bokeh
- 论文/Paper: http://arxiv.org/pdf/2208.12343
- 代码/Code: None
DSR: Towards Drone Image Super-Resolution
- 论文/Paper: http://arxiv.org/pdf/2208.12327
- 代码/Code: https://github.com/ivrl/dsr
Learning Continuous Implicit Representation for Near-Periodic Patterns
- 论文/Paper: http://arxiv.org/pdf/2208.12278
- 代码/Code: None
NeuralSI: Structural Parameter Identification in Nonlinear Dynamical Systems
- 论文/Paper: http://arxiv.org/pdf/2208.12771
- 代码/Code: https://github.com/human-analysis/neural-structural-identification
Take One Gram of Neural Features, Get Enhanced Group Robustness
- 论文/Paper: http://arxiv.org/pdf/2208.12625
- 代码/Code: None