Sparse-view reconstruction

[YSYYU]

Thank you for your excellent work! From your paper, I can see that there is still a section on sparse-view reconstruction. Where is the code for this section?

[justimyhxu]

We don't provide a test example for sparse-view reconstruciton.
But you can refer this function

GRM/test.py

Line 219 in 4cd9c39

def images2gaussian(images, c2ws, fxfycxcy, model, gs_path, video_path, mesh_path=None, fuse_mesh=False, optimize_texture=False):

to setup your input images and poses for reconstruction.

[YSYYU]

ok, thank you for your reply! May I ask which areas should be changed?

[justimyhxu]

You can refer this code block to set up your image and camera pose inputs.

GRM/test.py

Line 453 in 4cd9c39

c2ws, fxfycxcy = generate_input_camera(2.7, [[20, 225+30], [20, 225+150], [20, 225+270], [-10, 225+330]], fov=50)

[YSYYU]

Thanks your reply. So what do the parameters r=2.7 and pose [20, 225+30] represent respectively, and how are these parameters determined？