The FAQ[0] about the images has some interesting things in it:
> Why are there images of different sizes?
> The Cassini cameras are 1-megapixel cameras. A normal image is 1024 x 1024 pixels. Using a technique called "summation" the cameras have the ability to combine pixels together to get smaller but less noisy images. This results in smaller images that take a lot less time to readout out and take up less data volume. Summation is very useful if a scientist needs to conserve both. In the 2 x 2 mode, the camera takes a 2 x 2 pixel square and averages those values into a single pixel. Images in this mode will be 512 x 512 pixels. In the 4 x 4 mode, the camera takes a 4 x 4 pixel square and makes that a single pixel. Images in this mode are 256 x 256 in size.
> Why are there images of different sizes?
> The Cassini cameras are 1-megapixel cameras. A normal image is 1024 x 1024 pixels. Using a technique called "summation" the cameras have the ability to combine pixels together to get smaller but less noisy images. This results in smaller images that take a lot less time to readout out and take up less data volume. Summation is very useful if a scientist needs to conserve both. In the 2 x 2 mode, the camera takes a 2 x 2 pixel square and averages those values into a single pixel. Images in this mode will be 512 x 512 pixels. In the 4 x 4 mode, the camera takes a 4 x 4 pixel square and makes that a single pixel. Images in this mode are 256 x 256 in size.
[0] https://saturn.jpl.nasa.gov/faq/#raw_images