Efficient Multimodal Geometry for Road-Scene Perception: Curb Detection and Zero-Shot Depth

Abstract

Road-Scene Geometry has become a test of how well researchers can connect performance with evidence quality, resource limits, and transfer across settings. This article critically maps combining cross-sensor attention and efficient state-space decoding for transferable geometric perception. The source set brings together 2 focal papers with 13 independently retrieved publications whose DOI or publisher records were checked before inclusion. The analysis is organized around sensor fusion, depth priors, attention efficiency, domain transfer, and road safety. To avoid reading outcomes from different settings as equivalent, the review compares research framing, method assumptions, and test envelope. Across the literature, the evidence indicates that advances in road-scene geometry become credible when representation, objective, and evaluation protocol are evaluated together and when uncertainty about distribution shift is reported explicitly. This organization relates method selection to decision risk and exposes recurring transfer threats, and proposes a research agenda centered on clear comparators, adversarial conditions, and inspectable evidence.