Parallel smoothing

Parallel smoothing data, a part of algorithm transparency, is computed from equalized image during feature extraction by smoothing pixels along smoothed ridge orientation in all primary blocks that are within mask. Parallel smoothing data is used together with orthogonal smoothing to produce binarized image.

Visualization

The following image can be reconstructed from this stage's binary data. Visualization itself is not part of transparency data.

Format

Parallel smoothing data has the same format as equalized image. Pixel values are still in range -1 (white) to 1 (black), but pixels outside the mask are zero.

Example: 016-parallel-smoothing.cbor

{
  "width": 388,
  "height": 374,
  "cells": [
    0.0,
    0.0,
    "... skipped 145,108 items in range -0.999 to 0.980 ...",
    0.0,
    0.0
  ]
}