Web Interface for Interactive Face Reconstruction
Description: Existing systems for mental face reconstruction such as CG-GAN () require programming knowledge to set up the necessary environment and are therefore impossible to use for non-expert humans. Additionally, these systems are not designed to easily add new functionalities developed as research in this field grows. Therefore, the goal of this thesis is to re-implement this system as a website with a Python backend where the models are running. All the actions users perform on the website should be saved into an SQL database. Additionally, we will extend the functionality from previous work, and conduct a user study to evaluate different design choices of the website and optimise them for usability.
Supervisor: Florian Strohm
Distribution: 70% Implementation, 30% User study and evaluation
Literature:  Zaltron, Nicola, Luisa Zurlo, and Sebastian Risi. 2020. CG-GAN: An interactive evolutionary gan-based approach for facial composite generation. Proceedings of the AAAI Conference on Artificial Intelligence 34.