ZipSplat: Fewer Gaussians, Better Splats
For researchers in novel view synthesis and 3D reconstruction, ZipSplat provides a more efficient representation that adapts to scene complexity without retraining, setting new benchmarks in pose-free settings.
ZipSplat introduces a token-based feed-forward model for 3D Gaussian Splatting that decouples Gaussian placement from pixel grids, using k-means clustering to compress visual tokens into a compact set. It achieves state-of-the-art results on DL3DV and RealEstate10K with ~6× fewer Gaussians than pixel-aligned methods, surpassing the best pose-free baseline by 2.1dB and 1.2dB PSNR respectively.
Feed-forward 3D Gaussian Splatting methods reconstruct a scene from posed or pose-free images in a single forward pass, yet current approaches predict one Gaussian per input pixel, tying the representation budget to camera resolution rather than scene complexity. A flat wall and a richly textured object thus produce equally many Gaussians despite very different geometric needs. We propose ZipSplat, a token-based feed-forward model that decouples Gaussian placement from the pixel grid. A multi-view backbone extracts dense visual tokens, and k-means clustering compresses them into a compact set of scene tokens. Cross- and self-attention refine these tokens, and a lightweight MLP decodes each into a group of Gaussians with unconstrained 3D positions. Because clustering is applied at inference, a single trained model spans the quality-efficiency curve without retraining. ZipSplat operates without ground-truth poses or intrinsics, yet sets a new state of the art on DL3DV and RealEstate10K with ${\sim}6{\times}$ fewer Gaussians than pixel-aligned methods, surpassing the best pose-free baseline by 2.1dB and 1.2dB PSNR, respectively. It further generalizes zero-shot to Mip-NeRF360 and ScanNet++, outperforming all comparable baselines. Our project page is at ${\href{https://veichta.com/zipsplat}{https://veichta.com/zipsplat}}$.