Is each pixel then just part of the ray that gets shot through each circle?
How do these slides combine with the later slides about sub-aperture arrays? My understanding from those slides is that each pixel in one microlens captures the same image at a slightly different location, combining with these slides, does that mean that each microlens captures one subpart of the entire image, and then within each microlens you have the same small subpart captured at different angles in each (u, v) coord. And then you combine all the microlens together to get the overall image. Or does it work differently?
^ I think that's correct. The microlens behaves kind of like an array of cameras. The microlens' spatial location in the image corresponds to an (x, y) position on the sensor's surface. Each (u, v) point is supposed to correspond to a ray inside the camera from the sensor plane at this (x, y) point through the (u, v) point on the lens to the focal plane, which is how we get the overall disk image for each microlens.