September news update: Deformetrica v4.1 and MICCAI

Deformetrica 4.1 is out, MICCAI, here we come !

Deformetrica version 4.1 is out and this release is a major one. We will be presenting this new version at MICCAI in Granada, Spain on the 20th of September.

Here are a few changes that have been made to enable an easier use of the application:

  • the command-line has been updated to differentiate the estimate and compute features
  • a new python API is now available
  • a simple user interface has been implemented
  • many other improvements

We recommend using Conda to install Deformetrica:

conda install -c pytorch -c conda-forge -c anaconda -c aramislab deformetrica

 

Feel free to start a discussion on our Google group or open an issue on our Gitlab repository.

CLI commands: estimate and compute

The command-line application has been splitted into 2 distinct commands: estimate and compute.

Usage examples:

$ deformetrica estimate [--parameters parameters.xml] model.xml dataset.xml

$ deformetrica compute [--parameters parameters.xml] model.xml

Python API

A new python API is now available for python 3.5 and 3.6. This new feature makes it possible to embed Deformetrica within your application and scripts. This API opens Deformetrica to a variety of new usage scenarios.

from deformetrica.api import Deformetrica

deformetrica = Deformetrica(output_dir='output')
deformetrica.estimate_deterministic_atlas(...)

Graphical User Interface

A GUI has been developed to help with the different parameter configuration and a live graph has been added to help visualise the algorithm’s convergence.

Once installed, the GUI can be launched using the deformetrica gui command.

Keep in mind that this GUI is still in an alpha state and only the Deterministic Atlas feature is currently available.

MICCAI 2018 in Granada, Spain

We will be at MICCAI in Granada in Spain from the 16th to the 20th of September to present Deformetrica 4.1 at the ShapeMI workshop.

Here is an extract from the MICCAI 2018 Conference Book:

You can find our article here.

Feel free to come and talk to us !