This is an old revision of the document!

Welcome to BacTrack Wiki pages

BacTrack is an open-source multi platform java application intended to assist the researcher in tracking various organisms in sequences of images and obtaining and visualizing their quantitative data.

Motivation: A common task in today’s synthetic and systems biology studies is the analysis of various features of organisms using movies obtained through time-lapse microscopy.
This process consists of identifying organisms on a large number of images and measuring their features whilst keeping track of their lineage which is a very cumbersome and error prone task for the researcher.
To facilitate the analysis, various mostly automated tools and algorithms were developed in recent years but all of them targeting rather specific organisms and/or laboratory setups, since a general purpose automated solution to this detection problem is extremely hard if not impossible.

Solution: We've designed bactrack as a “platform”, that is as a collection of somewhat independent well defined modules working together. Such setup facilitates maintenance, customization and upgrades. This property is particularly useful for the tracking algorithms - in BacTrack it is possible to include multiple tracking algorithms (“best-of-breed approach”) and thus use the best fitted algoritm for the problem at hand.
It is even possible to mix-and-match algorithms in a way to

BacTrack uses a semi-automated approach: it relies on the researcher to define the initial cells and all cell divisions. Other than that, the application tracks cells using a newly developed algorithm described here. At any point, a researcher can intervene and correct any unsatisfactory cell detections. We’ve tested on four different organisms and obtained promising results. The application enables configurable lineage tree visualization and data export for use in third party programs.

Unlike most approaches (REFS here), our aim is to build a general-purpose solution for tracking and analysis of various organisms. To that end, it is necessary to build a multi-platform, user-friendly application that is open to changes and additions. Furthermore, we believe it is important to enable ways for researchers to collaborate and easily exchange their analysis, so that they can reproduce other’s results and/or take on a different perspective on existing, already processed data (movies). With that in mind, we’ve designed the application to be in some ways similar to movie editing software (e.g. LightWorks, Pinnacle Studio, etc.) where the user is given full command of every step of the detection process (Fig 1). User can set or modify selections at any frame. For this purpose application leverages the ImageJ program - an excellent open-source program well known to the biology community and heavily used in image processing and analysis. But unlike others, we’ve taken the source code of ImageJ and incorporated it into our project, so that we can retain full control of our graphical user interface and program behaviourbehavior. / /

QR Code
QR Code start (generated for current page)