3D Point-Cloud Learning And Road-Scene Geometry beyond Nominal Performance: Mechanisms, Uncertainty, and Deployment

Abstract

A central challenge in 3D point-cloud learning and road-scene geometry is to compare studies whose mechanisms and validation settings do not share a single denominator. The present review uses hierarchical spatial state-space aggregation for point-cloud classification and attention-based fusion of LiDAR and camera cues for curb detection as focal cases for a boundary-aware synthesis. The analysis combines two focal publications with 12 previously verified sources and organizes the evidence around neighborhood construction, hierarchy, state-space mixing, sampling robustness, and efficiency. Rather than pooling incompatible outcomes, it compares research questions, representations, controls, and validation envelopes. Across the evidence base, the decisive issue is alignment: neighborhood construction shapes what is observed, hierarchy shapes how it is compared, and efficiency governs whether the conclusion can be transferred. Uncertainty is most informative when reported as part of the result rather than treated as a postscript. The resulting framework supports reproducible comparison while preserving differences between study designs, and it identifies concrete points at which transfer claims should be narrowed or retested. The article-specific emphasis on mechanisms, uncertainty, and deployment is used to connect neighborhood construction with efficiency while keeping the two focal research settings analytically distinct.