Copyright (c) 2006 - 2012 Christoph Sommer, Michael Held & Daniel Gerlich
Gerlich Lab, IMBA Vienna, Austria. CellCognition is distributed under the term of LGPL License.

www.cellcognition.org
www.doc.cellcognition.org

To compile the ccore extension you need to adopt the library/include paths in the setup.cfg accordingly.

Dependcies are:

• libvigraimpex
• libtiff
• liblzma (only if libtiff is statically linked)

Remove the build- and dist directories and also the file cecog/ccore/_cecog.so(pyd)

python setup.py build_ext --inplace

python setup.py install --prefix=<path-to-prefix>

run the make file.

run build_win64_bin.bat

Additionally run build_helper\windows_sdk_env.bat before running build_win64_bin.bat.

The CecogAnalyzer package comes with batteries included.

It contains

- a small set of raw images (10 timepoints of H2b-aTubulin)
- the two classifiers for H2b and aTubulin to test classification
- a pre-configured settings file which is loaded on start-up.

You can

- test Object Detection of the primary (H2b) and secondary (aTubulin)
  channels
- retrain and test the classifier for H2b and aTubulin in Classification
- test the tracking and select events in Tracking (only six tracks are found
  within the 10 frames)
- for Error correction you need to install the R-project (see below)

The package contains a sub-folder Data with

- Settings
    - demo_settings.conf, the settings file which is loaded on startup
    - graph_primary.txt, an example for a graph definition file (H2b)
    - graph_secondary.txt, an example for a graph definition file
      (Tubulin)
    - position_labels.txt, position labels such as OligoID or GeneSymbol

- Classifier
    - the class definition and sample annotations to pick samples with the
      larger data set, feature and SVM models to test (or train) the H2b
      and aTubulin classifiers

- Images
    - the input folder of the raw images

- Analysis
    - the output folder where results are written to

With the included raw images picking of classifier samples is not possible since not all necessary positions/timepoints are included. Please download the larger H2b-Tubulin data.

With the included data and settings only six mitotic events with four frames duration are selected.

To perform motif selection and error correction as presented in our paper more timepoints are needed than the package contains. Larger data sets can be found online at downloads.

You also might want to increase the length of the selected tracks, especially after the pro-prometa onset. Increase therefore the values in Tracking -> Timepoints [post] and Timepoints [pre].