HIP Summer School

Philipp Heuser (HIP/DESY) , Sara Krause-Solberg (HIP Helmholtz Imaging Platform, DESY)

The Helmholtz Imaging Platform invites young researchers from all Helmholtz Centers for the 2021 (virtual) HIP Summer School. This summer school will provide on three consecutive Tuesdays an comprehensive introduction to the basics of scientific imaging and image handling (21.09.2021), to the use of Fiji/ImageJ (28.09.2021), and to artificial intelligence usage in the context of imaging (05.10.2021). Each day will be organised as mixture of lectures, tutorials and hands-on sessions. Since the contents build on each other, the days cannot be booked individually.

You can apply for participation below, we will distribute the places as equally as possible between the different centers. Applicants from one of the HIP projects will be given priority.

Registration is open until 03.09.2021!


- knowledge of Python

- python 3 installed on your computer

- colab (you will need a google account for this one. If you do not have one, or do not want one, please let us know.)


Imaging 101 (Dominik Jüstel, Helmholtz Zentrum München)

21.09.2021, 9:45-13:00 and 14:00-17:00

The first day of the HIP summer school will focus on the fundamentals of imaging: What are images? How do different imaging modalities generate imaging data? What are the physical and mathematical principles behind imaging?

During the first part of the day, representative imaging modalities and their physics will be discussed. Moreover, we will explore mathematical concepts and tools to simulate these modalities. In hands-on programming exercises you will implement a simple simulation of a modality of your choice guided by the tutors.

The second part revolves about the inverse imaging problem, i.e., how to generate images from the data acquired with imaging systems. After an overview of different methods, ranging from analytic solutions and optimization methods to probabilistic and data-driven approaches, you will implement a reconstruction algorithm for the data you simulated in the first part, again, with the assistance of the tutors.

Image Data Visualization using Fiji & Friends (Robert Haase - TU Dresden, Kyle Harrington - HIP/MDC)

28.09.2021, 9:45-13:00 and 14:00-17:00

During this day, we will explore capabilities of the ImageJ / Fiji ecosystem for scientific image data visualization and processing. Fiji is a very interactive software allowing to optimize image data presentation manually and thus, special emphasis will be put on reproducibility and good scientific practice. We will explore basic image filtering and segmentation techniques as well as advanced techniques for interactive image data deconvolution. Finally, an introduction to image processing automation will be given and we will integrate Fiji into more complex python based image analysis workflows using pyimagej.

AI & Imaging (Dagmar Kainmüller, MDC)

05.10.2021, 9:45-13:00 and 14:00-17:00

The last day of the HIP summer school is dedicated to deep learning for image analysis. The course will introduce convolutional neural networks, and will cover essential rules for designing your own networks, in particular when dealing with large image data. You will get hands-on experience in setting up and training your own networks for image analysis tasks like images classification and image segmentation. 



  • Tuesday, September 21
    • 9:45 AM 5:00 PM
      Imaging 101

      09:45 - 10:00 Check in
      10:00 - 10:15 General introduction
      10:15 - 10:55 Imaging 101. First part - Imaging modalities & physics
      10:55 - 11:00 break
      11:00 - 11:55 Imaging 101. First part - mathematics, modelling and simulation
      11:55 - 12:00 break
      12:00 - 13:00 Imaging 101. hands-on tutorial 1: simulation of an imaging modality
      13:00 - 14:00 lunch break
      14:00 - 14:55 Imaging 101. Second part - image reconstruction
      14:55 - 15:00 break
      15:00 - 15:55 Imaging 101. Second part - probabilistic viewpoint and data-driven methods
      15:55 - 16:00 break
      16:00 - 17:00 Imaging 101. hands-on tutorial 2: solving an image reconstruction problem

  • Tuesday, October 5
    • 9:45 AM 10:00 AM
      Setup 15m
    • 10:00 AM 11:30 AM
      A Practical Introduction to Machine Learning for Image Classification, and A Formal Introduction to Deep Learning 1h 30m
    • 11:30 AM 12:00 PM
      Exercise 1: Image Classification 30m
    • 12:00 PM 12:45 PM
      Break 45m
    • 12:45 PM 1:45 PM
      A Practical Introduction to Image Segmentation 1h
    • 1:45 PM 2:15 PM
      Exercise 2: Semantic Segmentation 30m
    • 2:15 PM 2:45 PM
      Exercise 3: Instance Segmentation 30m
    • 2:45 PM 3:30 PM
      Break 45m
    • 3:30 PM 3:50 PM
      How to Deal With Large Image Data 20m
    • 3:50 PM 4:10 PM
      Exercise 4: Tile-and-stitch 20m
    • 4:10 PM 4:50 PM
      Exercise 5: Instance Segmentation Challenge 40m
    • 4:50 PM 5:00 PM
      Discussion of Challenge Results 10m