Enhancing of Images

12th of March 2021

10:30-12:15/12:30

REGISTER

Note that this workshop will be completely online. The instructor will be online to answer all your questions and guide you through the theory part and the hands-on parts. No in-person presence is possible due to the COVID-19 pandemic.Agenda
10:30-11:30 - Hands-on session with guided exercises and examples (Python)
11:30-12:15/12:30 - Invited Talk (see below)

Invited Talk (speaker will present live)

Title: Deep learning based super-resolution of solar magnetic field images
Speaker: Anna Jungblut (see Team page) (Article from Anna Jungbluth, Xavier Gitiaux, Paul J. Wright, Shane A. Maloney, Carl Shneider, Alfredo Kalaitzis, Michel Deudon, Atılım Güneş Baydin, Yarin Gal, Andrès Muñoz-Jaramillo)

Abstract: Over the last 50 years, a variety of instruments have obtained images of the Sun’s magnetic field (magnetograms) to study its origin and evolution. Improvements in instrumentation allow us to image the Sun with increasing resolution, exposing more and more detailed structures. 

Unfortunately, differences in resolution, noise, and saturation characteristics make direct comparison of magnetograms taken by different instruments challenging. This poses a significant constraint for research applications that require high-resolution and homogenous data spanning time frames longer than the lifetime of a single instrument.

To address this issue, we use deep-learning to calibrate and super-resolve magnetograms taken by the Michelson Doppler Imager (MDI), and the Global Oscillation Network Group (GONG), to the characteristics of images taken by the Helioseismic and Magnetic Imager (HMI). HMI images are the highest resolution full-disk magnetograms available to date and are a factor of 4-5 larger than images taken by MDI and GONG.

Our network is trained on a subset of overlapping data between MDI (or GONG) and HMI. Since super-resolution is an ill-posed problem, many high-resolution images can explain the same low-resolution input. We constrain the solution space by adding additional terms into the loss function to improve the model’s ability to capture the physics of the Sun’ magnetic field.

Technical PrerequisitesYou will need to have access to a reasonably fast internet connection to be able to follow the live stream of the lecture. The exercises will be all carried out in Google Colab, therefore you only need Google Chrome installed on your computer. A google account is also necessary.Know-how PrerequisitesTo be able to follow this lecture, you will need an intermediate understanding of mathematics, linear algebra and statistics. No previous knowledge of neural networks is assumed. An intermediate Python experience will be required to be able to follow and work on the exercises.