Some questions about the SE3 motion field

[ZYCheng777]

When I read the code, I found that the size of the SE3 motion field is consistent with the input image size (For example, the input image size is 640192, and the size of the SE3 motion field is consistent with it). I would like to ask if it can be converted into a transformation matrix in the traditional sense (a 44 transformation matrix including translation and rotation).