DESIGNATOR: : a Tool for Automated GAN-based Testing of DNNs in Martian Environments

This project provides a GUI-based testing framework for evaluating the robustness and accuracy of Deep Neural Networks (DNNs) using a high-fidelity Mars simulator built with Unreal Engine. The simulator is embedded directly into the GUI, allowing researchers and practitioners to visually observe how segmentation models perform under a variety of simulated terrain and lighting conditions.