Stream Dictionary work in progress

If the user wanted to customizing rule set or neural networks for either the Color Particle Analyzer or for the letter matcher it demands making a categorizer setup file or programming.

In order to do this you need to understand what is in the different lazy data streams. This page is a dictionary of them.

Stream Dictionary in polygons

  • aspectRatio Aspect ratio of polygon, higher is wider, smaller is taller, 1 is even
  • curveArchCount A line can either be straight or be a curve arch
  • endPointCount Number of end points in a polygon
  • endpointBottomPointCount Number of end point is the bottom (half?)
  • endpointBottomRightThirdPointCount Number of end point is the bottom right third
  • endpointTopLeftThirdPointCount Number of end points in top left third
  • hardPointCount A point can either hard corner or a soft point this is the count of hard corners
  • holeCount Number of holes enclosed by a polygon
  • horizontalLineCount Number of roughly horizontal lines in a polygon
  • lineCount Number of line segments in a polygon
  • multiLineCount A multi line is made up of 2 or more points, but with no Y junctions
  • pointCount Number of points in a polygon
  • tJunctionCount Number of T junction in a polygon
  • tJunctionLeftCount T junction in the left half of the bounding box
  • tJunctionRightCount T junction in the right half of the bounding box
  • uJunctionCount Number of undecided junctions that can be T junction or Y junction
  • verticalLineCount Number of roughly vertical lines in a polygon

Tricky streams

multiLineCount

A multi line is made up of 2 or more points, but with no T or Y junctions. This is an implementation dependent property, more than a pure geometric property.

The letter J would be traced into one piece and would have multiLineCount == 1.

The letter I San-serif would be traced into one piece, but since it is a straight line that would be translated to a simple 2 point line and would have multiLineCount == 0.

The letter Y would be split at a T or Y junction and would be split into 3 pieces, but they could possibly be simple lines.