KIT Career ServiceStudierendeAbschlussarbeiten

Designing a Training Dataset for Neural-Operator Fragment Removal in Helium CT: Statistical Sufficiency and Physics-Aware Augmentation

Forschungsthema/Bereich
Medical Imaging for Modeling and Simulation
Typ der Abschlussarbeit
Master
Startzeitpunkt
-
Bewerbungsschluss
30.04.2027
Dauer der Arbeit
-

Beschreibung

Motivation
Helium computed tomography is an experimental imaging technique that provides high-resolution, low-dose images for various applications in ion beam therapy, including treatment planning, verifying patient positioning,
and managing motion. Recently, the concept of simultaneous helium imaging during carbon ion radiotherapy (CIRT) was proposed, offering new possibilities for real-time adaptive CIRT.

However, this approach presents significant challenges due to the presence of high-energy fragments generated by the helium beam. These fragments are detected by the imaging system, which degrades image quality and limits the benefits of helium-based imaging. As these fragments have comparable energies to primary ions but unpredictable trajectories, standard filtering techniques are often ineffective. In this context, deep learning (DL) approaches offer a promising solution.

Student Project
The performance of learned fragment removal is limited by the quality of its training data: this thesis first determines how many primary particles per calibration spot are needed to characterize the fragment contribution reliably, and then extends the dataset through physics-aware detector augmentation and discretization variation, so that neural operators generalise across detector responses and grid resolutions.

Voraussetzung

Voraussetzungen an Studierende
  • General knowledge in Computer Science and Biomedical Engineering.

Studiengangsbereiche
  • Ingenieurwissenschaften
    Bioingenieurwesen
    Elektrotechnik & Informationstechnik
    Informatik
    Biomedical Engineering
    Computer Science
    Medizintechnik


Betreuung

Titel, Vorname, Name
M. Sc. Beatriz Isabel Paquim do Cruzeiro
Organisationseinheit
Institut für Biomedizinische Technik
E-Mail Adresse
beatriz.cruzeiro@kit.edu
Link zur eigenen Homepage/Personenseite
Website

Bewerbung per E-Mail

Bewerbungsunterlagen
  • Anschreiben
  • Lebenslauf
  • Notenauszug

E-Mail Adresse für die Bewerbung
Senden Sie die oben genannten Bewerbungsunterlagen bitte per Mail an beatriz.cruzeiro@kit.edu


Zurück