ilastik - the Interactive Learning and Segmentation Toolkit
ilastik is a simple, user-friendly tool for image classification and segmentation in up to three spatial and one spectral dimension. Using it requires no experience in image processing.
ilastik has a convenient mouse interface for labeling an arbitrary number of classes in the images. These labels, along with a set of generic (nonlinear) image features, are then used to train a Random Forest classifier. In the interactive training mode, ilastik provides real-time feedback of the current classifier predictions and thus allows for targeted training and overall reduced labeling time. In addition, an uncertainty measure can guide the user to ambiguous regions of the data. Once the classifier has been trained on a representative subset of the data, it can be exported and used to automatically process a very large number of images.
The features are computed in the full 2D/3D/4D pixel neighborhoods, depending on the available data. While the provided set of features includes popular color, edge and texture descriptors, the plug-in functionality allows advanced users to add their own problem-specific features. Feature computation and classifier prediction are multi-threaded and fully exploit modern multi-core machines.
So far, we have used ilastik successfully on applications from the neurosciences (segmentation of EM images), systems biology (high throughput screening experiments) and industrial quality control.
This video illustrates neuron boundary classification in 3D isotropic volumes. Many more tutorial videos here!
News
- 2011-11-30 - New Windows versions of ilastik 05.06 (rc5) online. Besides some hotfixes the batch processing has improved.
- 2011-11-16 - Want to become a core developer? Janelia farm is funding a scientific programmer to help take ilastik to the next level. If you are interested, please apply to the ilastik mailing list, and include a sample of the most complex piece of C++ code that you have written in the past
- 2011-10-10 - ilastik 0.5.06 for Mac Os X now online (including several hotfixes)
- 2011-09-09 - First prototype of ilastik 0.6 now works as it should, and allows processing of "infinitely" large data sets.
- 2011-09-01 - Interactive segmentation ("carving") will be presented at MICCAI 2011 on September 18th. Code and data (courtesy of Graham Knott) will be released on this site.
- 2011-07-07 - We moved the ilastik sources from gitorious to github. Enjoy!
- 2011-05-05 - ilastik 0.6 is taking shape, with a fundamentally revamped architecture built around an innovative graph structure. Join our mailing list to make sure not to miss the release.
- 2011-02-13 Do you want to join the team and shape the future of ilastik? We have two open positions.
- 2011-01-24 - An "advertising flier" for ilastik 0.5 has been accepted at ISBI2011.
- 2011-01-24 - A variant of ilastik 0.5 with some minor extensions for synapse detection is described in this ISBI 2011 conference paper.
- 2010-12-15 - ilastik 0.5.05 now online (including several hotfixes)
- 2010-12-01 - Many new videos! Watch them here.
- 2010-11-22 - ilastik 0.5 for MacOSX has just been released. See Download page.
- 2010-10-27 - ilastik 0.5 has been released. See Download page for more details.
Main features of ilastik
- Support of up to 4D (3D multi-spectral) data
- Learning of multiple classes
- Generic features cope with a wide array of local image structures and textures
- Features can be computed in all dimensions
- Batch mode allows automated processing of new images of the same kind
- Multi-threaded execution reduces processing time
- No programming expertise needed to interactively explore images and train a classifier
- Plug-in mechanism offers extensibility
- Open source project, BSD license
- Friendly open source developer community :-)
Limitations:
- ilastik bases its predictions on local properties. Problems that require global context or relative configurations of parts cannot be solved with ilastik.
- ilastik does not make coffee (yet)