Fionagh Thomson       11:56 AM
Really engaging talk. Do you think PSF (point spread function) has the same meaning in Astro (telescopes) as in biomedicine (microscopes).

Marc HUERTAS-COMPANY       12:00 PM
thank you. i am not familiar with microscope imaging, but my feeling is yes. It is just the impulse response of the system.



Ian McCheyne       12:01 PM
How do you know that the fake data generated by the GAN is represnetative of the real data you don’t want to consider (because it is an artifact or for some other reason)? Or have i misi=understood the use of the GAN?

Marc HUERTAS-COMPANY       12:05 PM
Not sure I understand your question. The WGAN is trained on real data, so if the generative model has converged, the generated data are representative,  by construction. The main assumption is that rare objects in the observed dataset will not be properly generated