The Australian Optical Society 


  • Home
  • Imaging and Optical Physics seminar - Zoom webinar

Imaging and Optical Physics seminar - Zoom webinar

  • 02 Apr 2020
  • 3:00 PM
  • Zoom Meeting ID: 224 053 850

Join from a PC, Mac, iPad, iPhone or Android device:

Please click this URL to start or join. https://monash.zoom.us/j/224053850
Or, go to https://monash.zoom.us/join and enter meeting ID: 224 053 850

Join from a dial-in phone:

Dial: +61 3 99059666 (+61 3 9905 ZOOM) or +61 2 8015 2088
Toll-free dial: 877 369 0926 (US/Canada only)
Meeting ID: 224 053 850
International numbers available: https://monash.zoom.us/u/abcA8IGbXj


Webinar agenda:

1. Kaye Morgan and David Paganin (Monash University)

Describing Phase Contrast and Dark-field X-ray Imaging with the Fokker-Planck Equation

Abstract:

In recent years, methods of x-ray imaging have been developed to non-invasively capture three measurements of a sample; attenuation, phase and dark-field. Attenuation reveals dense structures like the bones, phase effects can reveal weakly-attenuating structures like the airways, and dark-field can reveal the presence of sub-pixel structures like the air sacs in the lungs. In some cases, it can be difficult to ‘untangle’ these three signals and dark-field effects are sometimes neglected. To better understand the interplay between these measurements, we extend the transport of intensity equation (TIE) to now include dark-field effects.  The resulting equation is known in other fields of research as the ‘Fokker-Planck Equation’, used to model coherent transport and diffusive transport of energy. We justify use of this equation in x-ray optics, show how it can model edge and curvature effects and suggest possible future applications.

 

2. Juan Nunez-Iglesias (Monash University)

Visualisation and analysis of large, n-dimensional images in the scientific Python ecosystem

Abstract:

Over the past 15 years, Python has grown from a niche programming language in science to the de-facto standard for much scientific analysis. In large part, the libraries NumPy, SciPy, pandas, and matplotlib are responsible for this rise. Together, they formed a standard core around which a rich ecosystem of interoperable libraries, such as scikit-image and scikit-learn, could grow. Now, the increasing number of bigger-than-RAM image datasets, as well as the deep learning revolution, have again fragmented the array-oriented computing landscape in Python (think of dask arrays, zarr, PyTorch, etc). I will describe efforts to unify these libraries under a common framework, and how that translates into large image visualisation and analysis for users of these libraries.


Powered by Wild Apricot Membership Software