发表机构
Shanghai Jiao Tong University; Jishu Technology Co., Ltd.（上海交通大学; 集度科技有限公司）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对3D高斯溅射模型基元过多导致成本高的问题，提出G$^2$ARD-GS几何引导蒸馏方法，在MatrixCity数据集上实现最优压缩性能，提升了外观适配与配准精度。
AI 中文摘要
密集彩色激光雷达地图能提供精确的城市级几何信息，但将其转化为3D高斯溅射（3DGS）时会保留数百万个基元，导致生成的模型存储、传输、渲染和适配成本高昂。激进的基元缩减可缓解该负担，但会移除稳定新视图合成及下游几何应用所需的局部表面支撑。我们提出G$^2$ARD-GS，一种几何引导的蒸馏方法，可将密集高斯先验（可实例化为免训练点云提升或训练好的GS模型）转化为紧凑可复用的表示。G$^2$ARD-GS逐步将先验整合为感知表面的代表，随后在构建时锚点约束下恢复固定拓扑上的外观，恢复过程中不添加或移除任何基元。在有限监督下，几何感知视图选择分配可用的视图预算。在MatrixCity数据集上，G$^2$ARD-GS在匹配的5倍至30倍压缩预算下，实现了最佳的峰值信噪比（PSNR）、结构相似性指数（SSIM）和学习感知图像块相似度（LPIPS），在PSNR上较PUP提升了3.2至6.8分贝。当作为冻结几何复用为紧凑模型时，其在轨迹外外观适配上较PUP 3D-GS提升了3.7至4.9分贝，且在30倍压缩下保留了剑桥国王学院的图像到模型配准精度。项目页面：this https URL。
英文摘要
Dense colored LiDAR maps provide accurate city-scale geometry, but lifting them into 3D Gaussian Splatting (3DGS) retains millions of primitives, making the resulting models costly to store, transmit, render, and adapt. Aggressive primitive reduction alleviates this burden, but can remove the local surface support needed for stable novel-view synthesis and downstream geometric use. We introduce G$^2$ARD-GS, a geometry-guided distillation method that converts a dense Gaussian prior instantiated either as a training-free point-cloud lift or a trained GS model into a compact, reusable representation. G$^2$ARD-GS progressively consolidates the prior into surface-aware representatives, then recovers appearance on the resulting fixed topology under construction-time anchor constraints, with no primitives added or removed during recovery. Under limited supervision, geometry-aware view selection allocates the available view budget. On MatrixCity, G$^2$ARD-GS achieves the best PSNR, SSIM, and LPIPS across matched $5\times$--$30\times$ compression budgets, outperforming PUP by $3.2$--$6.8$,dB in PSNR. When reused as frozen geometry, the compact model improves off-trajectory appearance adaptation by $3.7$--$4.9$,dB over PUP 3D-GS and preserves image-to-model registration accuracy on Cambridge KingsCollege at $30\times$ compression. Project page: https://patrick1159.github.io/gardGS-page/.
