While 3D Gaussian Splatting (3DGS) enables real-time rendering, its training demands workstation-level compute and memory, making mobile deployment impractical under minute-scale time budgets and limited peak memory. We present PocketGS, a mobile scene modeling paradigm that enables on-device 3DGS training under these tightly coupled constraints while preserving high-fidelity reconstruction. PocketGS resolves the fundamental tension between training efficiency, memory compactness, and modeling quality through three co-designed operators: $\mathcal{G}$ builds geometry-faithful point-cloud priors; $\mathcal{I}$ injects local surface statistics to seed anisotropic Gaussians, thereby reducing early conditioning gaps; and $\mathcal{T}$ unrolls alpha compositing with cached intermediates and index-mapped gradient scattering for stable mobile backpropagation. Extensive experiments demonstrate that PocketGS outperforms the powerful mainstream workstation 3DGS baseline under mobile budgets, delivering high-quality reconstructions and enabling a fully on-device, practical capture-to-rendering workflow.

翻译：摘要：尽管三维高斯泼溅（3DGS）技术能够实现实时渲染，但其训练过程仍需要工作站级别的计算与内存资源，使得在分钟级时间预算和有限峰值内存约束下难以实现移动端部署。我们提出PocketGS——一种在严格资源耦合约束下实现移动端3DGS训练的轻量化场景建模范式，同时保持高保真重建质量。通过三个协同设计的算子，PocketGS解决了训练效率、内存紧凑性与建模质量之间的根本性矛盾：$\mathcal{G}$算子构建几何保真的点云先验；$\mathcal{I}$算子注入局部表面统计信息以初始化各向异性高斯体，从而缩小早期条件差距；$\mathcal{T}$算子通过缓存中间结果的阿尔法合成与索引映射梯度散射，实现稳定的移动端反向传播。大量实验表明，在移动端预算约束下，PocketGS的性能超越主流工作站级3DGS基线，不仅实现高质量场景重建，还支持完整的移动端“采集-渲染”实用工作流。