Monty.Frameworks.Models.Salience.Strategies.Vocus2.MapCombine behaviour (ExTbpMonty.Core v0.1.0)

Copy Markdown View Source

How several feature maps become one salience map.

Mirrors tbp/monty/frameworks/models/salience/strategies/vocus2/vocus2.py::MapCombine, a Protocol in Monty so a config can name how its channels are combined. It is a behaviour here: named maps in — "L", "a", "b", "depth", "orientation" — and one map of the same shape out.

Monty's plain functions are map_max/1, map_sum/1, map_mean/1 in Monty.Frameworks.Models.Salience.Strategies.Vocus2; the weighted ones, which need configuration, are WeightedSum and WeightedMean.

Summary

Types

maps()

One map per feature, all the same shape.

t()

Callbacks

call(t, maps)

The maps reduced to one, pixel by pixel.

Types

maps()

@type maps() :: %{optional(String.t()) => [[float()]]}

One map per feature, all the same shape.

t()

@type t() :: struct()

Callbacks

call(t, maps)

@callback call(t(), maps()) :: [[float()]]

The maps reduced to one, pixel by pixel.