Can I use this method for lytro style light field?

[DarwinSenior]

As stated, since colmap fails to register for Lytro style data, I am trying to manually generate poses and mpi_bds. However, I think I might do something wrong here.

I figured out that the poses is a 5x3 matrix which generates the homography.
If the horizontal baseline and vertical baseline of each view is x and y and the image is of size hxw and the focus is f, is this the right matrix for poses?

1 0 0 (-y) h
0 1 0 (-x) w
0 0 1 0    f

[rodrygojose]

Hi @DarwinSenior ,
You should be able to use lytro-style photos if you manage to create the poses_bounds.npy file.
Please, notice that the stored cameras represent camera-to-world transformations and that the camera coordinate system is [ down, right, backwards ].
For more information, please check this previous issue:
#10

[DarwinSenior]

I was trying with CG generated images, but it seems the images wrapped together but the inference does not have any effects. In the code, I generate MPI with

ibr_runner = DeepIBR()
ibr_runner.load_graph('./checkpoints/papermodel/checkpoint')
mpis = run_inference(imgs, poses, mpi_bds, ibr_runner, num_planes, patched)

and did

    plt.imsave(os.path.join(savedir, 'mpi{:02d}/disps.png'.format(i)), mpis[i].disps)

for each image
I have got the following images from MPI. Is it possible to speculate where it got wrong?

[rodrygojose]

Are you using only 4 input images?
If that's the case I would recommend to use more. At least 9 (a grid of 3x3).
The line:
plt.imsave(os.path.join(savedir, 'mpi{:02d}/disps.png'.format(i)), mpis[i].disps)
does not save MPIs, but a visualization of the estimated disparity.

[DarwinSenior]

Thank you. I have solved it. It was my mistake in using too few planes as parameters and mixing up with num_planes and N

[bishengderen]

如上所述，由于 colmap 无法注册 Lytro 样式数据，因此我正在尝试手动生成姿势并mpi_bds。但是，我认为我可能会在这里做错事。

我发现这些姿势是一个5x3矩阵，它产生单调性。如果每个视图的水平基线和垂直基线为 x 和 y，图像大小为 hxw，焦点为 f，那么这是姿势的正确矩阵吗？

1 0 0 (-y) h
0 1 0 (-x) w
0 0 1 0    f

Hello, I would like to ask you how you generate the attitude matrix from lytro data and convert it into an NPY file?