Automatic enumeration of cell colonies with OpenCFU

Quentin Geissmann
qgeissmann@gmail.com


A digital version of this poster is available online. 
http://opencfu.sourceforge.net/extra/20130910-Poster.svg


Backround:
Assessing the number of cell in solution is essential in many fields of cell biology.
This is very often done manually by the experiementers.
Manual counting is time counsuming and subjective.
Software tools are not adopted by the community.


In many microbiology and cellular biology experiments, getting an accurate estimate of the number of viable cells in solution is critical.
A very common approach involves spreading a known volume of a cell suspension on an solid media and enumerate the number of colonies after growth.
The counting is often performed manually. This task is time consuming and raises concerns about reproducibility.
During the last decade, a few authors have targeted this question and developed solutions, but none seem to have been adopted.

[FIG1: Borring counting]

Aims:
Design an efficient algorithm to identify cell colonies on images.
Develop a free and open-source software implementing this method in a user friendly interface.

Methods:

[FIG2: pipeline]
[FIG3: Random forest]

The algorithm is implemented in C++ using OpenCV API. [OpenCV logo]
In addition to a command line interface, a modern GUI was designed using the Gtkmm API. [Gtk+ logo]


RESULTS:

Accurate
Fast
Robust
Usable






