Smooth convolutional filters instead of feature maps?

[mrgloom]

From the paper:

How do the learned convolutional filters change? Our proposed change smooths the internal feature maps for purposes of downsampling. How does training with this layer affect the learned convolutional layers? ... the anti-aliased networks (red-purple) actually learn smoother filters throughout the network, relative to the baseline (black). Adding in more aggressive low-pass filtering further decreases the TV (increasing smoothness). This indicates that our method actually induces a smoother feature extractor overall.

I wonder is it possible to smooth convolutional filters itself instead of feature maps?

[richzhang]

While that would help, it would not be antialiasing the exact signal that is to be downsampled, since there are subsequent nonlinearities, such as relus and maxpool.