PERF: Performant, Explicit Radiance Fields

We present a novel way of approaching image-based 3D reconstruction based on radiance fields. The problem volumetric is formulated as non-linear least-squares and solved explicitly without the use neural networks. This enables solvers with higher rate convergence than what typically used for networks, fewer iterations are required until convergence. volume represented using grid voxels, scene surrounded by hierarchy environment maps. makes it possible to get clean reconstructions 360° scenes where foreground background separated. A number synthetic real from well-known benchmark-suites successfully reconstructed quality par state-of-the-art methods, but at significantly reduced times.