Desktop Point Cloud Visualization Application

Partnered with a leading research organization to build a desktop visualization tool for exploring a fruit fly brain connectome — a dataset of millions of 3D points representing mapped neurons. Researchers needed a way to navigate, label, and modify massive point clouds that existing tools couldn't handle. Built a data pipeline with a Python API for programmatic control, Unity as the rendering engine, and Potree for efficient level-of-detail point cloud rendering. Implemented features including dynamic point addition/removal, persistent labels, depth lighting, eye-dome lighting, and interpolation. Exceeded the initial 10K point target by processing 100K+ points with sub-second responsiveness, and successfully rendered 10M points at 30+ FPS. UBC Engineering Physics capstone project (ENPH 459).