Master's Thesis @IDEA Lab

The successful applicant will be part of the Image Data Exploration and Analysis (IDEA) lab @AIBE FAU, which is the research group of Prof. Bernhard Kainz. We have pioneered new methodological developments in the field of medical image analysis and machine learning as evident from several international awards, prizes, and best paper awards. The group is internationally renowned for its research in medical deep learning, which has been proven by publications at top conferences (MICCAI, CVPR, ECCV) and in prestigious journals (IEEE TMI, Medical Image Analysis, npj Nature Digital Health, The Lancet, etc.).

Description

Machine learning applications in video frame analysis have shown promise in automatically identifying specific frames or detecting frames containing particular objects. However, one challenge that emerges is the sheer volume of predictions, often running into thousands for a single video. From a human-centric perspective, only the most pivotal frames are usually relevant. Despite the importance of determining the informativeness of an image, the issue remains an unresolved challenge in the computational domain. Furthermore, conventional clustering techniques often prove insufficient due to the non-Euclidean nature of image-embedding manifolds.

This project aims to address the challenge of selecting representative subsets of frames. The selected subset should encapsulate the essential variability and information required to effectively train a downstream machine learning model. For instance, a model trained for object detection on this subset should exhibit a performance analogous to one trained on the entire dataset.

Qualifications

Necessary qualifications:

  • Profound knowledge in Computer Science, Medical Engineering or Electrical Engineering (or similar study programmes)
  • Profound (practical) knowledge of the programming language Python
  • Profound (practical) knowledge of Machine learning
  • High motivation and high interest in the described topic
  • Organised and independent way of working with a focus on quality and accuracy
  • Very good communication skills in English

Desirable qualifications:

  • if available, please provide a list of your public GitHub repositories in your CV
  • if available, please add an employer's reference or credentials
  • if possible, please add the transcripts of records from FAU study programme

Supplementary description

Befristetes Forschungsvorhaben

Applications with missing information will be considered incomplete and will not be processed. Kindly refrain from sending emails regarding application status, we will contact only the successful applicant for this position.

The Department AIBE and FAU see itself as a progressive and forward-thinking employer. We welcome your application regardless of your age, gender, cultural and social background, religion, belief, disability or sexual identity. Friedrich Alexander University promotes professional equality for minority groups and women. These groups are therefore expressly encouraged to apply. Severely disabled persons within the meaning of the Severely Disabled Persons Act will be given preferential consideration in the case of equal professional qualifications and personal suitability if the advertised position is suitable for severely disabled persons. At the applicant's request, the Equal Opportunity Officer may be called in for the interview without any disadvantage to the applicant.

Notice

Für alle Stellenausschreibungen gilt: Die Friedrich-Alexander-Universität fördert die berufliche Gleichstellung der Frauen. Frauen werden deshalb ausdrücklich aufgefordert, sich zu bewerben.

Schwerbehinderte im Sinne des Schwerbehindertengesetzes werden bei gleicher fachlicher Qualifikation und persönlicher Eignung bevorzugt berücksichtigt, wenn die ausgeschriebene Stelle sich für Schwerbehinderte eignet. Details dazu finden Sie in der jeweiligen Ausschreibung unter dem Punkt "Bemerkungen".

Bei Wunsch der Bewerberin, des Bewerbers, kann die Gleichstellungsbeauftragte zum Bewerbungsgespräch hinzugezogen werden, ohne dass der Bewerberin, dem Bewerber dadurch Nachteile entstehen.

Ausgeschriebene Stellen sind grundsätzlich teilzeitfähig, es sei denn, im Ausschreibungstext erfolgt ein anderweitiger Hinweis.

Release date: 08.01.2024