Does `step_blur()` need a prefix like the other steps or is it a different kind of step?

[EmilHvitfeldt]

[corybrunson]

@EmilHvitfeldt i'd welcome your input. Currently, the _pd and _vpd conventions encode output formats (persistence diagrams or vectorized persistence diagrams). In most cases, this differs from the input format. (The exception is step_pd_degree(), which both inputs and outputs persistence diagrams.)

In contrast, step_blur() isn't intrinsically connected to persistent homology; it introduces blur to numerical arrays, with the output having the same dimensions as the input. If ≥2 additional steps are implemented that modify arrays understood as images, i would be tempted to create a convention for them, e.g. step_raster_blur().