Introduction

The woolz image processing library was developed for the purpose of microscope slide scanning. It is unique in the way in which the image strutures and file format define the notion of "domain", i.e. the spatial extent, which can be an arbitrary disconnected collection of pixel/voxels, is maintained independently of the values associated with each pixel or voxel location. This provides a very compact representation of image data, even without compression, and a very efficient mechanism set (union, intersection etc.), morphological (dilation, erosion etc.) and other binary operations (e.g. distance, skeleton). It also provides for efficient memory usage because a single representation of the image pixel values can be utilised (pointed at in C structure terms) by any number of independent domains. A good example is the segmentation of an image. Each independent segment or labelled region will have its own domain but can share the same values array, thereby avoiding copying and duplication, gaining both in processing time and memory usage. In addition to pixel or "grey-level" images, woolz also includes many other image related data structures, e.g. vertices, polylines, boundaries, histograms and complex 2D and 3D model strucutures (lines and surfaces).

The domains are coded using an interval notation akin to run-length encoding. Each domain has a bounding box

Introduction

Woolz Structures

Example Usage

Coding Conventions