Comments (3)
Our axis convention is unique and unfortunately doesn't share the classical opencv/openGL conventions. You could however utilize the following helper function to convert it to NeRF i.e. OpenGL format:
def convert_pose_PD_to_NeRF(C2W):
flip_axes = np.array([[1,0,0,0],
[0,0,-1,0],
[0,1,0,0],
[0,0,0,1]])
C2W = np.matmul(C2W, flip_axes)
return C2W
Optionally before converting it to OpenGL, you could consider canonicalizing the poses so they are centered at origin
c2w[:3, 3] = c2w[:3, 3] - obj_location
Thanks for your patience and we will release all of these details with our codebase soon!
from neo-360.
Correct, it is the center of the hemisphere used to generate camera poses. averaging the x/y axes values should also work the same way. For overfitting it probably shouldn't matter as much but for generalizability this should remove any biases in camera locations in world coordinates.
Other than this, it is a good practice to do pca on the poses which should give you similar results.
from neo-360.
Ok I see, that's what I suspected. Is there a reason for this choice?
Optionally before converting it to OpenGL, you could consider canonicalizing the poses so they are centered at origin
What is obj_location
, I saw it in the pose.json
file, the center of the hemisphere used to generate the camera views ?
I have been setting the world origin as the average of the cameras position on the x and y axes and keeping the z-axis value as 0.
Thank you for your availability :)
from neo-360.
Related Issues (13)
- Code Release of Neo-360 HOT 6
- Intrinsics missing HOT 2
- Dataset License? HOT 1
- Cameras in ground or objects HOT 1
- How to handle multi view images ? HOT 1
- Why Tri-plane matters instead of 3D feature grid?
- When do you plan to release the code?
- How to mitigate the CUDA-error: Out-of-memory? HOT 9
- Can Neo360 synthesize novel view from frontal camera images HOT 1
- Segmentation classes HOT 4
- Real world results from KITTI-360 evaluation HOT 1
- Is it possible to use image with unknown intrinsics and extrinsics , just initialised using DUSt3R to do the 3d reconstruction ? HOT 2
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 neo-360.