Hybrid Latents: Geometry-Appearance-Aware Surfel Splatting
For researchers in novel-view synthesis, this work improves geometry reconstruction and rendering efficiency by reducing the entanglement between geometry and appearance in Gaussian splatting models.
The paper introduces a hybrid Gaussian-hash-grid radiance representation for 2D Gaussian scene models that decouples geometry and appearance via frequency decomposition, achieving superior reconstruction fidelity with an order of magnitude fewer primitives compared to state-of-the-art Gaussian-based novel-view synthesis methods.
We introduce a hybrid Gaussian-hash-grid radiance representation for reconstructing 2D Gaussian scene models from multi-view images. Similar to NeST splatting, our approach reduces the entanglement between geometry and appearance common in NeRF-based models, but adds per-Gaussian latent features alongside hash-grid features to bias the optimizer toward a separation of low- and high-frequency scene components. This explicit frequency-based decomposition reduces the tendency of high-frequency texture to compensate for geometric errors. Encouraging Gaussians with hard opacity falloffs further strengthens the separation between geometry and appearance, improving both geometry reconstruction and rendering efficiency. Finally, probabilistic pruning combined with a sparsity-inducing BCE opacity loss allows redundant Gaussians to be turned off, yielding a minimal set of Gaussians sufficient to represent the scene. Using both synthetic and real-world datasets, we compare against the state of the art in Gaussian-based novel-view synthesis and demonstrate superior reconstruction fidelity with an order of magnitude fewer primitives.