Comments (9)
Hi, thanks for your interests.
1: The meaning of 'obj_corners_3d' is the eight corners of the 3D boxes that tightly bound the object in the camera coordinate space. 'obj_rest_corners_3d' is the position of these eight corners when the object is in the rest pose (i.e., the translation and rotation are zero). These two stuff is not important, and you could simply igore them. For the object, I only predict the object translation relative to the hand wrist.
2. Yes, I also don't use corner losses since it didn't bring me better performance. You could simply ignore it and set this loss to zero.
3. This code is in the final version, but it also contains some features (including predicting object corners) that I used for testing. For some features presented in the code but not mentioned in the paper, you could directly ignore them since I previously did experiments on them but didn't bring me gains.
from alignsdf.
Thank you for your reply in your busy schedule.
I want to ask one more question.
I find the final gt_obj_center is from the obj_corners_3d in function unpack_meta_params() in sdf_utils.py, is there any other way to get gt_obj_center(i use trimesh to compute the oriented box of .obj in mesh_obj, but the loss of center is not good)
best
xinkang
from alignsdf.
Hi Xinkang
Yes, I think you simply get the 'gt_obj_center' directly from the official obman annotation. When you opened one pkl file in the 'meta' directory, you could find a key called 'affine_transform', which is a 4x4 matrix. I remember the matrix[:3, 3] is the object center position in the camera coordinate system. You need to have a double check.
Hope it helps. I will also have a try to make the code easier to use at the end of this month.
Best,
Zerui
from alignsdf.
Hi Zerui
Thanks for your patient reply, I'm a beginner and appreciate your guidance. Your work is excellent, and I wish you better work and better articles in the future.
Best,
Xinkang
from alignsdf.
You are welcome :)
from alignsdf.
Hi,
I have put my preprocessed data at https://drive.google.com/drive/folders/1GoaA6vB6TwAAHmaobVo5GjRoCq2wT21R, which could help resolve your issue.
Best,
Zerui
from alignsdf.
Hi,
I have put my preprocessed data at https://drive.google.com/drive/folders/1GoaA6vB6TwAAHmaobVo5GjRoCq2wT21R, which could help resolve your issue.
Best, Zerui
Hi,
Thank you ,zerui.
Thank you for your data, I will use it properly.
By the way, I'm not familiar with the dexYCB dataset. Can you share your code for processing the dexYCB dataset and generating mesh.
Thank you for your selfless sharing.
Best,
Xinkang
from alignsdf.
Hi @zhangxinkang,
You could refer to #8 to have more tips about processing DexYCB dataset and generating meshes.
Best,
Zerui
from alignsdf.
Hi @zhangxinkang,
You could refer to #8 to have more tips about processing DexYCB dataset and generating meshes.
Best, Zerui
Hi, zerui, I am sorry i have not noticed the closed issue,thanks for your guidance.
from alignsdf.
Related Issues (20)
- 'obj_corners_3d' and 'obj_rest_corners_3d' in obman meta file HOT 4
- Missing test/mesh_obj_rest/ for obman test HOT 3
- How to generate dexycb split names? HOT 2
- Dex_YCB dataset organization HOT 2
- Question Regarding the SDF Sampling
- Question Regarding the SDF Scale Factor HOT 8
- Question about the training epochs HOT 4
- Issue related to DexYCB dataset cropping HOT 4
- Dataset download HOT 2
- occur "do not support renderer in this machine" problem HOT 2
- occur "do not support renderer in this machine" problem
- The difference between sdf_hand and sdf_hand_mini HOT 6
- Test the model on novel images
- **源 HOT 5
- Training HOT 3
- Issue related to the result of Ote HOT 2
- mesh_hand and mesh_obj folders HOT 1
- Pre-trained weights of AlignSDF HOT 1
- How to evaluate like Table 4 in ECCV paper? HOT 1
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 alignsdf.