Comments (4)
Hi,
we use them SMPL-X body model, but do not train hand pose and facial expression. This is why the output contains only global orientation, body pose, body shape, and translation. The 'missing' parameters (expression, hand, eye pose etc.) are set to mean pose (all zeros), so they're not part of the output.
The different formats for body_pose and global_rot are just conversions between rotation representations. I assume you're expecting axis angle rotation representations for global orientation and body pose. You can convert global_rot which is a rotation matrix (1,3,3) to smpl-x format which is in axis angle (1,3). raw_global_rot is a 6D rotation representation which we use for training (see this paper).
from shapy.
Thank you for your reply. If I want to get the full parameters, what tools I can use, for example, smplify-x? I am actually using a rgb-d cameras thus I can still provide depth information.
Update
I have successfully used smplify-x to get the SMPL-X parameters. However, the speed is extremely low... Is there any faster solution to get this? BTW, I found the translation parameter got back from smplify-x is "camera_translation". Is it the same as the "transl" used in shapy and the example I proposed in my first question?
from shapy.
Another question is related to the scale of shapy and smplify-x. In the follwing meshlab environment, I plot the mesh obtained from shapy (the leftmost), the mesh obtained from smplify-x (the rightmost), and the point cloud obtained from rgb-d image (the middle). As we can see, the mesh obtained from smplify-x lies in the origin. However, the mesh from shapy deviates a lot, in terms of z-axis. Moreover, both shapy and smplify-x deviate y-axis a little bit. How to solve this problem? Is it because shapy and smplify-x lacks of depth information?
from shapy.
@felixshing hey have you found the answer why we have a mismatch between camera params of shapy and smplify-x ? I am also having the same issue.
from shapy.
Related Issues (20)
- any way to do the opposite of regressor? HOT 1
- Running setup.py install for mesh-mesh-intersection ... error HOT 6
- Issue in testing video HOT 1
- Running regression demo.py in container with Ubuntu 18.04 HOT 1
- SHAPY predicts hugely differences in shape parameters between 2 close images. HOT 1
- A2S with ModelAgency data HOT 3
- Keypoint detection HOT 1
- Waist measurement problem HOT 1
- ModuleNotFoundError: No module named 'attributes.attributes_betas' HOT 1
- HOTO? HOT 5
- AttributeError: Caught AttributeError in DataLoader worker process 0. HOT 9
- Unable to install mesh_mesh_intersection
- Predictions Problems
- All meshes are same
- [Windows] Solved problem(mesh-mesh-intersection)
- For Attributes to Shape (A2S), I want to modify the input data and change it to input bust, waist, hip and height
- No image is created.
- Measurements accuracy
- Problem with Cuda Extension Generation on Colab
- Use multiple images for one output image
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 shapy.