I believe the architecture and training process should be very similar for both models, so we should work on them in parallel. Ideally we want the differences in their performance to demonstrate the difference between local protein info vs. regional/cell-wide protein shifts.
It seems like a pressing question is how we go about sampling regions of interest (ROI).
A stub for the two cnn baseline models.
I believe the architecture and training process should be very similar for both models, so we should work on them in parallel. Ideally we want the differences in their performance to demonstrate the difference between local protein info vs. regional/cell-wide protein shifts.
It seems like a pressing question is how we go about sampling regions of interest (ROI).
Tasks to be broken down into smaller issues: