SPIN: An Open Simulator of Realistic Spacecraft Navigation Imagery
SPIN is a simulator designed to generate spacecraft navigation imagery. Built on Unity, it is capable of generating detailed images of spacecrafts in different scenarios.
Research
My research specializes in the generation and exploitation of synthetic data for computer vision. I have developed data pipelines using game engines, procedural generation, and generative models to solve visual perception tasks where real-world data is limited or unavailable. This work spans several published papers and covers applications in semantic segmentation, pose estimation, visual odometry, and depth estimation.
Synthmantic LiDAR: A Synthetic Dataset for Semantic Segmentation on LiDAR Imaging
Synthmantic LiDAR is a synthetic dataset built using a modified version of the Carla simulator, designed to train and evaluate semantic segmentation models on LiDAR imaging. Our work was accepted into the 2024 International Conference on Image Processing (ICIP).
Using Semantic Segmentation to Boost Reinforcement Learning Performance
For my Master's thesis, I thought about how using semantic segmentation simplified certain environments for reinforcement learning agents, and built a system to research this idea using the classic Super Mario game. Our work was published in the Multimedia Tools and Applications journal.