Open snowflakewang opened 9 months ago
Thank you for your interest in our paper.
CCM is a Coordinate Map where each pixel value represents its coordinate in a normalized space. You can think of it as first-order spatial information, while a normal map in the world coordinate system represents second-order information.
Regarding the open-source release of our code, I apologize for not having more time to organize it due to my current workload. As some recent works have proved that introducing geometric information improves the quality of 3D generation(RichDreamer, GeoDream, etc.), and I have been busy with another project recently, so I have not been able to release the code the first time. However, I plan to rearrange the relevant code as soon as possible during the upcoming Spring Festival holiday (mid Feb.).
Thank you very much for your attention :D
Thank you for your reply! According to your explanations, I recognize the process as follows: when rendering a map from a camera pose, use (x,y,z) coordinates instead of (r,g,b) colors to compute the map.
Did I realize your idea correctly?
Hello, I am a novice student in 3D Generation from USTC. I am impressed by the geometry and texture shown on your project page. They are quite highly detailed! I have two questions: 1) When reading your paper, I am confused about the concept of the Canonical Coordinates Map (CCM). According to the paper, CCM is obtained by rendering canonically oriented and uniformly normalized objects. So, is it a kind of normal map? If not, how can I better realize this concept?
2) Do you have any plan to release the code, or further checkpoints?
Thanks a lot!! :)