Hello, I am amazed by the effect of your article. I hope you can help me solve a doubt. There are some data sets, such as renderme360, whose camera center is not half of the length and width of the image. How does 'prcppoint' solve this problem? I want to solve this problem by rewriting the projection matrix (the results of your paper are already very good, I rewrote it based on another article, and that article did not take into account the center offset)
But the result is not ideal
How do you use prcppoint? Or how do you rewrite diff-plane-rasterization?
Hello, I am amazed by the effect of your article. I hope you can help me solve a doubt. There are some data sets, such as renderme360, whose camera center is not half of the length and width of the image. How does 'prcppoint' solve this problem? I want to solve this problem by rewriting the projection matrix (the results of your paper are already very good, I rewrote it based on another article, and that article did not take into account the center offset) But the result is not ideal How do you use prcppoint? Or how do you rewrite diff-plane-rasterization?