# ZPressor: Bottleneck-Aware Compression for Scalable Feed-Forward 3DGS

arXiv: https://arxiv.org/abs/2505.23734 · HTML: https://mlsanity.com/paper/2505.23734/

## Key facts

- Cited by 32 papers in the map (data through 14 Oct 2026).
- Builds on 14 of its 24 references that are in the map.
- First posted on arXiv 29 May 2025.

## Builds on (most cited first, 14 of 14)

- [FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioning](https://mlsanity.com/paper/2307.08691/) (2023)
- [Direct Voxel Grid Optimization: Super-fast Convergence for Radiance Fields Reconstruction](https://mlsanity.com/paper/2111.11215/) (2021)
- [Training Deep Nets with Sublinear Memory Cost](https://mlsanity.com/paper/1604.06174/) (2016)
- [Depth-supervised NeRF: Fewer Views and Faster Training for Free](https://mlsanity.com/paper/2107.02791/) (2021)
- [Scaffold-GS: Structured 3D Gaussians for View-Adaptive Rendering](https://mlsanity.com/paper/2312.00109/) (2023)
- [MVSplat360: Feed-Forward 360 Scene Synthesis from Sparse Views](https://mlsanity.com/paper/2411.04924/) (2024)
- [MuRF: Multi-Baseline Radiance Fields](https://mlsanity.com/paper/2312.04565/) (2023)
- [R2L: Distilling Neural Radiance Field to Neural Light Field for Efficient Novel View Synthesis](https://mlsanity.com/paper/2203.17261/) (2022)
- [Explicit Correspondence Matching for Generalizable Neural Radiance Fields](https://mlsanity.com/paper/2304.12294/) (2023)
- [SelfSplat: Pose-Free and 3D Prior-Free Generalizable 3D Gaussian Splatting](https://mlsanity.com/paper/2411.17190/) (2024)
- [PixelGaussian: Generalizable 3D Gaussian Reconstruction from Arbitrary Views](https://mlsanity.com/paper/2410.18979/) (2024)
- [GauSS-MI: Gaussian Splatting Shannon Mutual Information for Active 3D Reconstruction](https://mlsanity.com/paper/2504.21067/) (2025)
- [Lightning NeRF: Efficient Hybrid Scene Representation for Autonomous Driving](https://mlsanity.com/paper/2403.05907/) (2024)
- [Discovering Common Information in Multi-view Data](https://mlsanity.com/paper/2406.15043/) (2024)

## Built on by (most cited first, 30 of 32)

- [VolSplat: Rethinking Feed-Forward 3D Gaussian Splatting with Voxel-Aligned Prediction](https://mlsanity.com/paper/2509.19297/) (2025)
- [C3G: Learning Compact 3D Representations with 2K Gaussians](https://mlsanity.com/paper/2512.04021/) (2025)
- [Latent Spatial Memory for Video World Models](https://mlsanity.com/paper/2606.09828/) (2026)
- [DriveGen3D: Boosting Feed-Forward Driving Scene Generation with Efficient Video Diffusion](https://mlsanity.com/paper/2510.15264/) (2025)
- [LongSplat: Online Generalizable 3D Gaussian Splatting from Long Sequence Images](https://mlsanity.com/paper/2507.16144/) (2025)
- [Revisiting Depth Representations for Feed-Forward 3D Gaussian Splatting](https://mlsanity.com/paper/2506.05327/) (2025)
- [GlobalSplat: Efficient Feed-Forward 3D Gaussian Splatting via Global Scene Tokens](https://mlsanity.com/paper/2604.15284/) (2026)
- [Off The Grid: Detection of Primitives for Feed-Forward 3D Gaussian Splatting](https://mlsanity.com/paper/2512.15508/) (2025)
- [World-R1: Reinforcing 3D Constraints for Text-to-Video Generation](https://mlsanity.com/paper/2604.24764/) (2026)
- [MeshSplat: Generalizable Sparse-View Surface Reconstruction via Gaussian Splatting](https://mlsanity.com/paper/2508.17811/) (2025)
- [Feed-Forward 3D Scene Modeling: A Problem-Driven Perspective](https://mlsanity.com/paper/2604.14025/) (2026)
- [AnyDepth: Depth Estimation Made Easy](https://mlsanity.com/paper/2601.02760/) (2026)
- [MotionWeaver: Holistic 4D-Anchored Framework for Multi-Humanoid Image Animation](https://mlsanity.com/paper/2602.13326/) (2026)
- [TriSplat: Simulation-Ready Feed-Forward 3D Scene Reconstruction](https://mlsanity.com/paper/2605.26115/) (2026)
- [UniMesh: Unifying 3D Mesh Understanding and Generation](https://mlsanity.com/paper/2604.17472/) (2026)
- [SplatWeaver: Learning to Allocate Gaussian Primitives for Generalizable Novel View Synthesis](https://mlsanity.com/paper/2605.07287/) (2026)
- [Block3D: Efficient Text-to-3D Generation via Block-Wise Diffusion](https://mlsanity.com/paper/2608.19567/) (2026)
- [CylinderSplat: 3D Gaussian Splatting with Cylindrical Triplanes for Panoramic Novel View Synthesis](https://mlsanity.com/paper/2603.05882/) (2026)
- [AirSplat: Alignment and Rating for Robust Feed-Forward 3D Gaussian Splatting](https://mlsanity.com/paper/2603.25129/) (2026)
- [Lite3R: A Model-Agnostic Framework for Efficient Feed-Forward 3D Reconstruction](https://mlsanity.com/paper/2605.11354/) (2026)
- [Learn2Splat: Extending the Horizon of Learned 3DGS Optimization](https://mlsanity.com/paper/2605.15760/) (2026)
- [ZipTok3D: High-Fidelity 3D Tokenization with Compact Token Prefixes](https://mlsanity.com/paper/2609.01740/) (2026)
- [ReCoSplat: Online Feed-Forward Gaussian Splatting via Render-and-Compare](https://mlsanity.com/paper/2603.09968/) (2026)
- [Last-Layer-Centric Feature Recombination: Unleashing 3D Geometric Knowledge in DINOv3 for Monocular Depth Estimation](https://mlsanity.com/paper/2604.26454/) (2026)
- [RoSplat: Robust Feed-Forward Pixel-wise Gaussian Splatting for Varying Input Views and High-Resolution Rendering](https://mlsanity.com/paper/2605.13093/) (2026)
- [Learning Stable Canonical Worlds for Novel View Synthesis and Beyond](https://mlsanity.com/paper/2606.23027/) (2026)
- [Compact Feed-Forward 3D Gaussians via Saliency-Guided Primitive Merging](https://mlsanity.com/paper/2608.10712/) (2026)
- [LocusGS: Spatially Grounded Tokens for Feed-Forward 3D Gaussian Splatting](https://mlsanity.com/paper/2608.12825/) (2026)
- [AVSplat: Dense-View Feed-Forward 3D Gaussian Splatting with Assist-View Preconditioning](https://mlsanity.com/paper/2609.05925/) (2026)
- [FeCoSplat: Feedback-Guided Compression for Feed-Forward 3D Gaussian Splatting](https://mlsanity.com/paper/2609.33330/) (2026)

## Abstract (arXiv)

Feed-forward 3D Gaussian Splatting (3DGS) models have recently emerged as a promising solution for novel view synthesis, enabling one-pass inference without the need for per-scene 3DGS optimization. However, their scalability is fundamentally constrained by the limited capacity of their models, leading to degraded performance or excessive memory consumption as the number of input views increases. In this work, we analyze feed-forward 3DGS frameworks through the lens of the Information Bottleneck principle and introduce ZPressor, a lightweight architecture-agnostic module that enables efficient compression of multi-view inputs into a compact latent state Z that retains essential scene information while discarding redundancy. Concretely, ZPressor enables existing feed-forward 3DGS models to scale to over 100 input views at 480P resolution on an 80GB GPU, by partitioning the views into anchor and support sets and using cross attention to compress the information from the support views into anchor views, forming the compressed latent state Z. We show that integrating ZPressor into several state-of-the-art feed-forward 3DGS models consistently improves performance under moderate input views and enhances robustness under dense view settings on two large-scale benchmarks DL3DV-10K and RealEstate10K. The video results, code and trained models are available on our project page: https://lhmd.top/zpressor.