VisSection: Style Transferring for Visualization in Anatomy Education

Type: 
DA
Persons: 
1

Description

Anatomy education provides medical students with an in-depth understanding of the anatomical structures, which often cannot be acquired with traditional means, due to a lack of resources. Computer support and advanced visualization techniques for medical volume data have the potential to enhance the education in medical disciplines. In order to achieve a high degree of realism, it is important to be able to provide a realistic visualization of a certain anatomical structure, similar to that of real tissues. Style transfer is the process of creating an image that simultaneously matches the content of an image and the style of another image. The goal is to preserve the arrangement of the content image and to represent it using the colors and the local structures from the style image.

This project aims to research the possibility of using interactive visualizations of 3D medical data for anatomy education, by simulating the physical models with the photorealistic rendering and the illustrations from books with the style transferring. The ability of switching between realistic visualization and illustrations while exploring the anatomical structures could help students to better understand them.

Tasks

The project we propose aims to research the possibility of using style transfer techniques in combination with photorealistic visualization of anatomical structures from medical volume data to increase insight and understanding of human anatomy. The expected results could be used to create interactive visualizations that allow the transition between photorealistic visualizations and illustrations, facilitating a better understanding of the anatomical structures.

Requirements

  • Knowledge of English language (source code comments and final report should be in English)
  • Interest and Knowledge in Medical Visualization, in particular Illustrative Visualization and Anatomical Education, as well as Machine Learning
  • Good programming skills
  • Basics in Machine Learning
  • Creativity and enthusiasm

Environment

The project should be implemented as a standalone application, desktop or web-based (to be discussed).

Contact

For more information please contact Renata Raidou (rraidou@cg.tuwien.ac.at).