AI 中文总结
TRACE-GS是首个利用训练时的特权几何条件，将扩散先验适配到稀疏视图3DGS恢复的在线策略轨迹蒸馏框架，在各设置下实现了一致性能提升与强泛化性。
AI 中文摘要
我们提出了TRACE-GS，这是一种在线策略轨迹蒸馏框架，在训练时利用特权几何条件，从而将扩散先验适配到稀疏视图3D高斯溅射（3DGS）恢复任务中。与追求日益复杂的恢复架构不同，我们发现现有基于扩散的方法存在一个更根本的局限：独立加噪状态下的监督未覆盖推理过程中达到的状态。在稀疏视图3DGS中，欠约束的几何结构从一开始就使去噪产生偏差，且由此产生的偏差会随过程推进不断累积。TRACE-GS则执行在线策略轨迹蒸馏：由额外训练视图提供的更丰富几何条件的教师模型，会为稀疏视图学生模型自身的推进过程提供目标，在每个访问状态下对齐去噪方向和跨视图响应。这种仅用于训练的几何条件使TRACE-GS处于利用特权信息学习（LUPI）的设置中。在部署时，仅保留稀疏视图学生模型，其恢复的渲染结果作为伪观测值用于3DGS优化。据我们所知，TRACE-GS是首个从特权几何中导出在线策略监督以用于稀疏视图3DGS恢复的方法，在各数据集和稀疏视图设置下均取得了一致的性能提升和出色的泛化能力。
英文摘要
We present TRACE-GS, an on-policy trajectory distillation framework that leverages privileged geometric conditioning at training time, thereby adapting a diffusion prior to sparse-view 3D Gaussian Splatting (3DGS) restoration. Rather than pursuing increasingly sophisticated restoration architectures, we identify a more fundamental limitation shared by existing diffusion-based approaches: supervision at independently noised states does not cover those reached during inference. In sparse-view 3DGS, under-constrained geometry biases denoising from the outset, and the resulting deviations compound along the rollout. TRACE-GS instead performs on-policy trajectory distillation: a teacher conditioned on richer geometry from additional training views supplies targets along the sparse-view student's own rollout, aligning denoising directions and cross-view responses at each visited state. This training-only geometry places TRACE-GS in the learning using privileged information (LUPI) setting. At deployment, only the sparse-view student is retained, and its restored renderings serve as pseudo-observations for 3DGS refinement. To the best of our knowledge, TRACE-GS is the first to derive on-policy supervision from privileged geometry for sparse-view 3DGS restoration, achieving consistent gains and strong generalization across datasets and sparse-view settings.
