Welcome to gensei’s documentation!

gensei = generator of serial images, or a karate style (gen = mysterious, sei = control, system), meaning that the methods of quantitative histology (or stereology) are often a mysterious process for us :)

web: http://code.google.com/p/gensei/

code: http://github.com/rc/gensei/

Get the latest sources:

$ git clone git://github.com/rc/gensei.git
$ cd gensei

and read the README for installation instructions.


The software is used for generating series of two-dimensional sections through a volume filled with ellipsoids, cylinders and fibres with a priori known statistical properties. Arbitrary series of images are produced in order to represent models of histological structures, such as cells and fibres of the extracellular matrix forming biological tissues. These images represent either serial consecutive sections or sections sampled in a systematic uniform manner. The user can set the number and types of classes of three-dimensional objects (ellipsoids, cylinders), dimensions of the reference volume, size of the generated images, the number of objects and sections to be generated, length-to-width ratio and volume fraction of the objects. Three perpendicular series of sections through the same reference volume and the same objects may be generated, thus testing the isotropy of the objects. The output of the simulation is stored as three series of image files together with a detailed report containing the statistical properties of the generated objects, i.e. number of objects per class, volume fraction of the objects within the reference volume, surface density of the objects, number of intersections between the objects and the sections, and rotation of the objects around uniquely determined spatial vectors. By comparing the true and section-based estimates of the volume, surface and length, the settings of various stereological grids used for estimation can be tested. The user can e.g. simulate the effect of section thickness on the stereological estimate of volume, surface and length. The resulting images have a high contrast so the objects can be segmented.


Indices and tables

Table Of Contents

Next topic

User’s Guide

This Page