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|谷歌 实时3D人体关键点检测&姿态估计,已加入mediapipe!
论文: https://arxiv.org/abs/2206.11678
代码:https://mediapipe.dev/
摘要: 我们展示了 BlazePose GHUM Holistic,这是一种用于 3D 人体地标和姿势估计的轻量级神经网络管道,专为实时设备上推理而定制。BlazePose GHUM Holistic 支持从单个 RGB 图像进行动作捕捉,包括头像控制、健身追踪和 AR/VR 效果。我们的主要贡献包括 i) 一种新的 3D 地面实况数据采集方法,ii) 更新的 3D 身体跟踪与额外的手部标志和 iii) 从单目图像估计全身姿势。
最新论文整理
ECCV2022
Updated on : 14 Sep 2022
total number : 6
StoryDALL-E: Adapting Pretrained Text-to-Image Transformers for Story Continuation
- 论文/Paper: http://arxiv.org/pdf/2209.06192
- 代码/Code: https://github.com/adymaharana/storydalle
Weakly-Supervised Stitching Network for Real-World Panoramic Image Generation
- 论文/Paper: http://arxiv.org/pdf/2209.05968
- 代码/Code: None
Check and Link: Pairwise Lesion Correspondence Guides Mammogram Mass Detection
- 论文/Paper: http://arxiv.org/pdf/2209.05809
- 代码/Code: None
PointScatter: Point Set Representation for Tubular Structure Extraction
- 论文/Paper: http://arxiv.org/pdf/2209.05774
- 代码/Code: https://github.com/zhangzhao2022/pointscatter
CenterFormer: Center-based Transformer for 3D Object Detection
- 论文/Paper: http://arxiv.org/pdf/2209.05588
- 代码/Code: https://github.com/tusimple/centerformer
Adversarial Coreset Selection for Efficient Robust Training
- 论文/Paper: http://arxiv.org/pdf/2209.05785
- 代码/Code: None