Comments (5)
We recorded the value in millimeter. You can justify easily the unit of our data by: if it is generally <10, then its in meter; if it is generally >100, then its in millimeter
As for origin, if you want to project 3D coordinates into 2D, please check the closed issue 5 and 6, and I also past the answer here too: The 3d keypoints in RGBto3D_train.json are actually R_cam * X_global, so you don't need R_cam but you need T_cam to project correctly into 2D.
The 2d keypoints are X=K_cam(R_cam * X_global + T_cam) then X[:, :, :2] /= X[:, :, 2:3] is the 2D coordinates
Otherwise, if you just want to do training and evaluation without 2D, then the position of the origin is not important, since our proposed evaluation are always pelvis aligned (global translation ignored) But the origin, according to the formula in second paragraph (The 3d keypoints in RGBto3D_train.json are actually R_cam * X_global), should be the same origin of world coordinate system
from wholebody3d.
Alternatively, you could also download H3WB dataset in a format commonly employed for 3D pose estimation tasks here.
We provide an accompanying data preparation class for this format. We also recommend this format for your experiments.
In this format all 3D keypoints are in global coordinate system.
from wholebody3d.
@Zhuyue0324 I want to use 3D_global, so I just need to figure out R_cam and then use it to divide x,y,z data. Am I correct?
from wholebody3d.
Yes, you are right. Or you just use the data which @nerminsamet has provided, which directly has 3D global coordinates
from wholebody3d.
Thank you all very much. Hava a nice day!
from wholebody3d.
Related Issues (20)
- code for how to construct the dataset from multi-view geometry ? HOT 1
- question about 3d coordinate system? HOT 10
- issue about data preparation script(data path issue) HOT 12
- Willing to perform motion forecasting HOT 3
- Facing Error with dimensions HOT 6
- Inference script HOT 1
- Which file, global_3d or camera_3d, should be used to train the model? HOT 5
- How to train the modle HOT 4
- Data loading HOT 1
- Module import HOT 1
- data normalization HOT 4
- NumPy data for test set HOT 1
- Can you provide the code for training `Large SimpleBaseline` for 3D whole body lift from complete 2D whole body keypoints (2D->3D)? HOT 3
- the visualize results is correct ? HOT 3
- The coordinates of keypoints_3d HOT 10
- Could you provide a script for projecting 3D keypoints from RGBto3D_train.json to 2D coordinate points of the image? HOT 3
- 关于测试集 HOT 1
- 关于数据集名称 HOT 3
- 您好,请问在 2Dto3D_trian_part1.json 文件中读取得到的三维关键点如何和S1的某一帧动作进行匹配呢 HOT 4
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from wholebody3d.