Package org.shapelogic.imageprocessing

Image processing: Segmentation, vectorizers and annotation

See:
Description

Interface Summary

IParticleCounter
IPixelTypeFinder	Interface that are used for several Vectorizer.
LazyPlugInFilter<E>	The idea is that you can call a plugin in both active and lazy mode.
PixelJumperByte	Not sure if I need this, but this is a little bit of structure build on top of the ByteProcessor.
SBPixelCompare	Interface for doing pixel compare.

Class Summary

AreaComparator	Compare area by pixel size.
BaseMaxDistanceVectorizer	Vectorizer that is splitting lines based on max distance to line between end points.
BaseParticleCounter	ParticleCounter count number of particles in a particle image.
BaseVectorizer	Input image needs to be binary, that is gray scale with inverted LUT.
ChainCodeHandler	Chain Code For MultiLine.
CircleFinder	Class that can find a circle in MultiLine.
DirectionBasedVectorizer	DirectionBasedVectorizer is a Vectorizer based on direction.
DistanceBasedColorHypothesisFinder	Find a color hypothesis based on just distance of colors.
LineProperties	LineProperties contains properties that are important for a line when you are dealing with curved multi line.
LineVectorizer	LineVectorizer is a vectorizer using short line of default length 5.
MaxDistanceVectorizer	Vectorizer that is splitting lines based on max distance to line between end points.
NeighborChecker	Neighbor Checker.
ParticleCounter	ParticleCounter count number of particles in a particle image.
PixelTypeCalculator	PixelTypeCalculator stores some values for points and calculated the type of points based on them.
PointProperties	LineProperties contains properties that are important for a point when you are dealing with curved multi line.
PriorityBasedPixelTypeFinder	Classify points to find out what type they are.
SBByteCompare	Compare implementations for gray scale.
SBColorCompare	Compare implementations for RGB.
SBPendingVertical	Contains information about one line that needs to be checked.
SBSegmentation	High level class for segmentation.
SBSimpleCompare	Abstract class fro compare.
SegmentCounter	Segmentation for 24 bit RGB and 8 bit Gray.
ShortLineBasedVectorizer	Base class for vectorizers that are using a local short line, to determine when to set point on multi line.
SimplePixelTypeFinder	Find the type of a pixel.
StreamVectorizer	Same vectorizer as MaxDistanceVectorizer, but logic implemented with streams.

Enum Summary

GeometricType	Not sure if I should just reuse PixelType.
PixelType	Enum with types for annotating pixels.

Package org.shapelogic.imageprocessing Description

Image processing: Segmentation, vectorizers and annotation

Content of image processing package

• Segmentation
• Vectorizers
• Annotation

Vectorizers

A vectorizer is an algorithm for transforming bitmap lines into vector lines.

There are 3 different implementations vectorizers in ShapeLogic.
There are currently 3 different vectorizer implementations. They all assume that you are working on binary gray scale images images. So before you can use a vectorizer you have to convert your bitmap to this format.
The 2 first had a lot of problems and are now deprecated.
The vectorizers are part of a class hierarchy of different vectorization algorithms, all sub classed from BaseVectorizer.

Common to current vectorizer implementations

• Assumes that you have a gray scale image with only black and white in it
• Are modifying the bitmap

It would probably be a good idea to loosen some of these restrictions.

MaxDistanceVectorizer

The main idea is that this will read a whole multi line at a time. Then later it will split it according to max distance of pixels to the line between start and end point of the multi line.
This vectorizer is the one that works well.

LineVectorizer

LineVectorizer is a vectorizer reading sequences of short line of length 5. If 2 consecutive lines are close in direction you can merge them. But if they are far away in angle you create a new line.
This approach ran into problems and has been deprecated for now.

DirectionBasedVectorizer

The main idea is that this will read a whole multi line at a time. It is using a priority of different types of points around an intersection to read trough an intersection where you can take different ways.
This approach ran into problems and has been deprecated for now.

Segmentation

Color and gray scale segmentation algorithm. The segmentation algorithm works on both RGB and 8 bit gray scale images.

It is currently writing out information to a dialog box.