发表机构
Xidian University（西安电子科技大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
提出多源协同训练与类别感知融合框架，利用数字孪生、扩散重绘等数据源，在DSVT上实现路侧LiDAR仿真到现实的3D检测，UrbanTwin V2X-Real测试集综合得分0.7421。
AI 中文摘要
弥合路侧LiDAR仿真到现实之间的差距需要解决多个耦合的差异，包括场景几何、采样密度、回波模式和行人尺度。本报告提出了一种用于Sim2Real 3D检测的多源协同训练和类别感知融合框架。该方法将数字孪生扫描、扩散重绘扫描、密度稳定扫描和行人形态对齐样本组织到一个具有互补角色的统一训练池中。在共同的DSVT检测框架内，源特化专家分支保留这些角色，同时优化相同的检测目标。在推理时，预定义的类别感知融合路径整合了几何稳定和校准感知分支用于车辆，采样互补分支用于卡车，以及形态一致证据用于行人。一种无标签的点云中心混合方法进一步细化了几何定位。在UrbanTwin V2X-Real隐藏测试集上，该统一系统取得了0.7421的综合得分，其中3D mAP@0.5为0.4518，真实感得分为0.8871。结果表明，数据源之间稳定、可解释的协作比无约束的模型输出聚合更有价值。
英文摘要
Bridging the simulation-to-reality gap in roadside LiDAR requires addressing several coupled discrepancies, including scene geometry, sampling density, return patterns, and pedestrian scale. This report presents a multi-source collaborative training and class-aware fusion framework for Sim2Real 3D detection. The method organizes digital-twin scans, diffusion-redrawn scans, density-stabilized scans, and pedestrian morphology-aligned samples into a unified training pool with complementary roles. Within a common DSVT detection formulation, source-specialized expert branches preserve those roles while optimizing for the same detection objective. At inference, a predefined class-aware fusion pathway integrates geometry-stable and calibration-aware branches for vehicles, sampling-complementary branches for trucks, and morphology-consistent evidence for pedestrians. A label-free point-cloud center blend then refines geometric localization. On the UrbanTwin V2X-Real hidden test set, the unified system achieves a combined score of 0.7421, with 3D mAP@0.5 of 0.4518 and a realism score of 0.8871. The results indicate that a stable, interpretable collaboration among data sources is more valuable than unconstrained aggregation of model outputs.
Comments7 pages,2 figures
