[Question] How to use more than 3 cameras?

[eliasab16]

It seems like the number of cameras is hardcoded to just 3 (and the docs only mention three, image0, image1, and image2). I have a few clarifying questions:

• Is this a deliberate architectural constraint?
• If extending to 4+ views is feasible, would it be as simple as adding "image3" to that loop (and retraining/fine-tuning)?
• If not, would stitching two camera views into a single image be a viable alternative, or would the resolution loss from resizing hurt more than it helps for precision tasks?