发表机构
Keio University; University of Stuttgart（庆应义塾大学; 斯图加特大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对单视图前馈3DGS重建的空间冗余问题，提出冻结基础模型的后处理剪枝与循环精化压缩流水线，实现高内存缩减并保持渲染质量。
AI 中文摘要
近年来的单视图前馈3D高斯泼溅（3DGS）生成方法为每条相机光线预测固定数量的高斯体，引入了严重的空间冗余。大多数现有的压缩策略针对多视图设置，利用跨视图一致性，因此与单图像模型不兼容。我们的见解是保持基础模型冻结，对生成的高斯体应用后处理剪枝和循环精化，而不是重新训练基础前馈网络以直接输出紧凑表示。因此，我们提出了一种与骨干网络无关的单视图前馈3DGS压缩流水线，该流水线将基于重要性评分的剪枝机制与可训练的轻量级循环精化模块相结合，迭代更新存活的图元以恢复图像质量。我们的结果表明，该方法能与现有基线无缝集成，同时保持新视角渲染保真度并实现高内存缩减。此外，我们的方法支持灵活的推理时保留比率，以满足应用需求。
英文摘要
Recent single-view feed-forward 3D Gaussian Splatting (3DGS) generation predicts a fixed number of Gaussians per camera ray, introducing severe spatial redundancy. Most existing compaction strategies target multi-view setups to exploit cross-view consistency and are incompatible with single-image models. Instead of retraining the base feed-forward network to directly output compact representations, our insight is to keep the base models frozen and apply post-hoc pruning and recurrent refinement to the generated Gaussians. Consequently, we propose a backbone-agnostic compaction pipeline for single-view feed-forward 3DGS that couples an importance-score-based pruning mechanism with a trainable, lightweight recurrent refinement module, which iteratively updates the surviving primitives to restore image quality. Our results demonstrate seamless integration with existing baselines while preserving novel-view rendering fidelity and achieving high memory reduction. Furthermore, our method supports flexible inference-time keep ratios for application needs.
