Lorentz Workshop “Quantitative 3D X-Ray Imaging:
from Tomographic Images to Metrics”, Jan. 11-15 2016, Lorentz@Oort.

Modern X-ray tomography techniques, combining a
mixture of advanced experimental imaging and novel mathematical reconstruction
methods, are a topic of intensive study where many advances are currently made.
After obtaining the 3D images, the next step forward in addressing scientific
questions is constituted by the determination of quantitative parameters
describing the phenomenon under investigation. The main
goal of this workshop was to initiate a thorough discussion about the
extraction of quantitative object features based on 3D X-ray imaging.

The workshop brought together around 50 participants
from different communities (computational, engineering, experimental,
metrology, biomedical, materials science) involved in X-ray tomography both
with synchrotron and conventional sources. For the discussions, three groups
were initially formed around the three key topics of

1.
Dimensional metrology

2. Quantification of dynamic processes

3.
Quantitative X-ray orientation imaging

The discussions in each of the groups developed in
very different way. The approach of dimensional metrology aims to define
guaranteed system properties for quantitative imaging, such that the full
imaging pipeline must be taken into account. As this topic was relatively new
to many of the mathematical and experimental imaging participants, much of the
discussion focused on becoming familiar with the basic metrological and
statistical concepts. The dynamic processes discussion group followed a more
hands-on approach, focusing the discussion on how to define useful 4D (3D plus
time) phantom objects that can be used for validating reconstruction algorithms
and on the challenge of visualization in dynamic tomography. Finally, the X-ray
orientation imaging group was relatively small, discussing mainly opportunities
to join software efforts in order to benefit as a community from each other’s
work.

The workshop featured a substantial number of
excellent, high profile speakers who gave a broad series of lectures about all
the concepts involved in quantitative tomography. During the workshop, we
gradually started to realize that the problems of carrying out quantitative
time-resolved X-ray imaging requires a very high degree of standardization,
which is difficult to achieve in practice. The key positive result of this
workshop is the awareness that has risen in the community of the difficult
steps that have to be taken to achieve this goal. As a concrete step, a new
software project was started during the workshop to develop an open-source
implementation of a 4D phantom object that can be used for standardized,
configurable simulations.

Overall, the workshop was received in a very
positive way. We expect that in the coming year, several publications will
emerge from the collaborations initiated at this Lorentz Workshop. This
workshop was partially funded by the EU EXTREMA COST Action, a network grant
for advanced tomography.