Hello, and thank you very much for your excellent work and open-sourced implementation.
We are currently attempting to apply your method to our own custom dataset. In our case, we have a set of images along with their corresponding intrinsic and extrinsic camera parameters. The extrinsic parameters are provided as transformation matrices from the camera coordinate system to the world coordinate system.
Could you kindly advise us on how we should organize or format this data in order to use it directly with your pipeline?
For example:
Are there any specific folder structures or naming conventions required?
In what format should the poses be saved (e.g., text files, numpy arrays)?
Are there any preprocessing steps we should be aware of?
We would really appreciate your guidance. Thank you again for your time and contribution to the community!
Hello, and thank you very much for your excellent work and open-sourced implementation.
We are currently attempting to apply your method to our own custom dataset. In our case, we have a set of images along with their corresponding intrinsic and extrinsic camera parameters. The extrinsic parameters are provided as transformation matrices from the camera coordinate system to the world coordinate system.
Could you kindly advise us on how we should organize or format this data in order to use it directly with your pipeline?
For example:
Are there any specific folder structures or naming conventions required?
In what format should the poses be saved (e.g., text files, numpy arrays)?
Are there any preprocessing steps we should be aware of?
We would really appreciate your guidance. Thank you again for your time and contribution to the community!