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：
标题：CLIFF: Carrying Location Information in Full Frames into Human Pose and Shape Estimation
代码：https://github.com/huawei-noah/noah-research/tree/master/CLIFF
论文：https://arxiv.org/abs/2208.00571
自顶向下的方法在3D人体姿势和形状估计领域占据主导地位，因为它们与人体检测分离，允许研究人员专注于核心问题。然而，裁剪是它们的第一步，从一开始就丢弃了位置信息，这使得它们无法在原始相机坐标系中准确预测全局旋转。为了解决这个问题，我们建议在这个任务中携带全帧位置信息（CLIFF）。具体来说，我们通过将裁剪的图像特征与其边界框信息连接起来，向CLIFF提供更全面的特征。我们在更宽的全帧视野下计算2D重投影损失，采用与在图像中投影的人相似的投影过程。CLIFF由全球位置感知信息提供并监督，它直接预测全球旋转以及更精确的关节姿势。此外，我们提出了一种基于CLIFF的伪地面真值注释器，它为野外二维数据集提供高质量的三维注释，并为基于回归的方法提供关键的全面监督。对流行基准测试的大量实验表明，CLIFF的表现明显优于现有技术，并在AGORA排行榜上排名第一（SMPL算法跟踪）。
最新论文整理
ECCV2022
Updated on : 20 Oct 2022
total number : 7
LaMAR: Benchmarking Localization and Mapping for Augmented Reality
- 论文/Paper: http://arxiv.org/pdf/2210.10770
- 代码/Code: None
GraphCSPN: Geometry-Aware Depth Completion via Dynamic GCNs
- 论文/Paper: http://arxiv.org/pdf/2210.10758
- 代码/Code: None
Attaining Class-level Forgetting in Pretrained Model using Few Samples
- 论文/Paper: http://arxiv.org/pdf/2210.10670
- 代码/Code: None
Depth Contrast: Self-Supervised Pretraining on 3DPM Images for Mining Material Classification
- 论文/Paper: http://arxiv.org/pdf/2210.10633
- 代码/Code: None
PoseGPT: Quantization-based 3D Human Motion Generation and Forecasting
- 论文/Paper: http://arxiv.org/pdf/2210.10542
- 代码/Code: None
PERI: Part Aware Emotion Recognition In The Wild
- 论文/Paper: http://arxiv.org/pdf/2210.10130
- 代码/Code: None
Anomaly Detection Requires Better Representations
- 论文/Paper: http://arxiv.org/pdf/2210.10773
- 代码/Code: None
CVPR2022
NeurIPS
Updated on : 20 Oct 2022
total number : 5
TOIST: Task Oriented Instance Segmentation Transformer with Noun-Pronoun Distillation
- 论文/Paper: http://arxiv.org/pdf/2210.10775
- 代码/Code: https://github.com/AIR-DISCOVER/TOIST
Learning to Discover and Detect Objects
- 论文/Paper: http://arxiv.org/pdf/2210.10774
- 代码/Code: None
CroCo: Self-Supervised Pre-training for 3D Vision Tasks by Cross-View Completion
- 论文/Paper: http://arxiv.org/pdf/2210.10716
- 代码/Code: None
Hierarchical Reinforcement Learning for Furniture Layout in Virtual Indoor Scenes
- 论文/Paper: http://arxiv.org/pdf/2210.10431
- 代码/Code: None
On the Adversarial Robustness of Mixture of Experts
- 论文/Paper: http://arxiv.org/pdf/2210.10253
- 代码/Code: None
