News |  Sitemap |  Contact
PDF Export  | 


Administrative Information


University of California Los Angeles

Presentation Information

Full / Half Time Slot: Half Time Slot (25 min)

Contact / Speaker Name

Albert Cardona

Presentation Title

TrakEM2: an ImageJ-based program for morphological data mining and 3D modeling

Participant Requirements

A general idea on electron microscopy and histology, and very basic notions of computer science.

Biography of Speaker

Neuroscience is my current field of research, my primary interest in the developmental and computational aspects of nervous system architecture. In year 2005 I completed a PhD on the embryonic development of planaria. Throughout my thesis work I developed and shared simple ImageJ plugins for serial reconstruction and three-dimensional modeling. This year I finally moved on to neurobiology of Drosophila, although I keep a close relationship with the Institute of Neuroinformatics, University/ETH Zurich, where we are developing a comprehensive software application for morphological data mining and serial reconstruction. In summary, I'm a biologist who's hobbies are converging into the research workflow.


The modern morphologist relies extensively on computer-aided image acquisition and analysis. While most of the desired functionality exists in separate software packages, the latter don't cross-talk well, and further, given the high volume of data a lab can generate in a very short time interval, keeping track of data and metadata becomes nearly impossible. Here we present our design of an ImageJ-based program aiming at providing an integral solution to morphological data mining. In particular, our program aims at storing, displaying and analyzing digital images, while aiding in extracting both three-dimensional and relational models of the sample. Furthermore, the program frees the researcher from tedious data-management tasks, and provides the means to deploy the data remotely over the internet. We have explored possible user interfaces with a prototype program, and found that a design directing the researcher from a self-provided abstract object ontology not only results in the most consistent data acquisition and valuable aid in modeling, but also self-constructs a rapid, light-weight means of accessing data through hierarchical trees and graphs.

© Luxembourg Institute of Science and Technology | Legal Notice