Open Torment123 opened 1 year ago
Hi, here is a detailed explation: http://www.songho.ca/opengl/gl_projectionmatrix.html.
Thanks @yzslab for the reference.
However, according to the derivation in the link (which I followed its derivation and think it's right) the P[2, 2] should be z_sign (zfar + znear) / (zfar - znear), whereas the code's implementation is z_sign zfar / (zfar - znear). Any ideas??
Possibly a potential bug have to double check my notes, but if you think about it we never use the normalized z value anywhere in the computations
On Mon, Oct 23, 2023, 09:48 Sungwon Hwang @.***> wrote:
Thanks @yzslab https://github.com/yzslab for the reference.
However, according to the derivation in the link (which I followed its derivation and think it's right) the P[2, 2] should be z_sign (zfar + z_near) / (zfar - znear), whereas the code's implementation is z_sign zfar / (zfar - znear). Any ideas??
— Reply to this email directly, view it on GitHub https://github.com/graphdeco-inria/gaussian-splatting/issues/376#issuecomment-1774613144, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACGXXYIHMPHDJ3LQZ7MRQ2DYAYONLAVCNFSM6AAAAAA6LJXQH6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTONZUGYYTGMJUGQ . You are receiving this because you are subscribed to this thread.Message ID: @.***>
Hi, I'm a bit new to multi-view geometry, could someone help me explain the getProjectionMatrix function in graphics_utils.py ? Is it calculating the camera intrinsic matrix? but there is no focal length?