ct3d/at3d Tutorial
Getting Started with ct3d and at3d
An Introduction to Tracking Cells Using the ct3d
tool
Contents
Introduction
Installation
First Example
Further Remarks
Introduction
ct3d is a command-line tool for tracking cells in
time series of 2D images or 3D image stacks. ct3d
performs the basic cell tracking as a command line
tool. Visualization, exploration, and evalutation of the results can
be done through the graphical user interface of at3d.
Installation
After downloading
ct3d, follow the installation instructions and make both
the ct3d and at3d command accessible in your path. For
Linux and MacOS, you can download compiled binaries. Note
that ct3d
requires libtiff as well
as lpsolve Version
5.5. at3d requires libtiff
and qt. If you
build ct3d/at3d yourself, you need full installations
including header files. If you download the binaries, the dynamic
libraries need to be accessible from the corresponding location in
your shell environment.
First Example
You
can download
a synthetically generated data set from the ct3d
homepage. Unpack ct3d_sample.tar.gz and change to
the ct3d_sample directory.
Launch ct3d through the
shell command.
After invocing ct3d, you can enter further ct3d commands. Typing
create -i ./input -o ./output -m 500 -M 5000 -s 200
|
will start computation of the component trees. After a few minutes,
you will be able to inspect the created component trees by typing
So far, you have only created component trees for all time frames in
your image sequence. To perform the actual cell tracking, type
and save your results using
After quitting ct3d by the
command, you can launch at3d. After launching, use the Open
button in the upper left corner to load the data from
the ct3d_sample/output folder by multi-selecting
all .tif files in this folder. You can now browse through
the tracking result, e.g. using the Next and Previous
button. You can select and eventually delete cells by clicking their
ID boxes in the panel on the right side of the window; cells can also
be filtered out using the Apply Filter button. Cells are
generally identified through their color. Assigning the same color to
two segments hence corrects for oversegmentation of a cell. Finally,
clicking the buttons in the Statistics box will provide
quantified output of cell motility parameters in textual form.
Preparing Data
Note that ct3d may only process uncompressed grayscale images. An easy way to convert a given sequence of tif images into suitable format is to use the mogrify command from the imagemagick software. If your tiff format images reside in the current working directory, simply type
mogrify *.tif -compress None -colorspace Gray
|
This can be easily adapted to convert from other data formats. Refer to the imagemagick documentation for details.
Further Remarks
Both ct3d and at3d support a number of
features not convered in this turotial. We only briefly mention those
for further self-exploration - for details, contact the authors.
ct3d allows to re-use previously computed
component trees using the load
command.
at3d provides numerous features for
filtering and selecting specific cells in tracking
results.
- Filtering is also supported
by
ct3d's filter command
- You can further automatize
ct3d by piping a text
file into the CT3D command. Simply list the commands you
would type in the ct3d command line, separated by line
breaks, in a text file, and pipe this into CT3D.
-
ct3d can also read color files. These will be
re-interpreted as grayvalue images through averaging the RGB
componentes. This can make sense for re-coloring existing results.
ct3d can be used in a very simple way
for computing cosegmentations for colocalization studies: Simply
create a folder containing only the two images to be co-segmented and
run ct3d as described in the example above.
ct3d/at3d Getting Started
Hang Xiao and Axel Mosig, Pattern
Discovery Group, CAS-MPG Partner Institute for Computational Biology,
Shanghai, China