Testing Datas

[liaojinli]

Would you like to ask how to test own datas?

[apple2373]

I also want to test the model on my own dataset. I tried but failed. It does not output anything looks like a face... It would be nice if there's some guidelines to apply the model on our own dataset. From what i have guessed, the following are what we should prepare.

intrinsic matrix K
[[f, 0, cx],
[0, f, cy],
[0, 0, 1]]
where f is the focal length, and cx and cy is the principal point of the image (usually W/2 and H/2 unless we crop images).

Rotation matrix R -- 3x3 array of world-to-camera rotation matrix representing the camera orientation.
Camera position T -- 3x1 array of camera position.

Then,
KRT is K.dot(R).dot(T)
rigid_mat is np.hstack([R,T])

I used COLMAP with this notebook to extract the K, R, and T. https://github.com/google/nerfies/blob/1a38512214cfa14286ef0561992cca9398265e13/notebooks/Nerfies_Capture_Processing.ipynb
So "The local camera coordinate system of an image is defined in a way that the X axis points to the right, the Y axis to the bottom, and the Z axis to the front as seen from the image."

Am I doing it right? I used 512x512 image, but the mesh i got is terrible. It's not even have a shape of face.