Unfolding 3D Gaussian Splatting via Iterative Gaussian Synopsis
Iterative Gaussian Synopsis is a novel framework for compact and progressive rendering of 3D Gaussian Splatting (3DGS) models. It uses a top-down 'unfolding' scheme, starting from a full model and iteratively creating coarser Levels of Detail (LODs) via learnable pruning. This method balances high visual quality across all levels with significant storage reduction, making 3DGS practical for resource-constrained applications.