Abstract

Biology has become a data-driven science. Much of the raw data now comes in the form of images, creating a need for easy-to-use tools for automated quantitative analysis.

In this talk, I will describe and demo our current best attempt at realizing such a tool: ilastik (http://ilastik.org) is a convenient tool for image classification and segmentation which requires no experience in image processing.

The interactive training of a powerful classifier allows to distinguish an arbitrary number of classes (such as different tissue, different organelles, etc) provided that these are distinguishable by local appearance. The program provides real-time feedback of the current classifier predictions and thus allows for targeted training and overall reduced labeling time. Once the classifier has been trained on a representative subset of the data, it can be used to automatically process a very large number of images. ilastik works on gray value, color or spectral images with up to three spatial dimensions.

I will demo selected applications from high-throughput screening, the neurosciences and mass spectrometric imaging, and be around to experiment with images supplied by the audience.

ilastik is open source and available from http://ilastik.org