DVPSFormer: Efficient Online Depth-aware Video Panoptic Segmentation for Autonomous Driving

Safe autonomous navigation requires a holistic understanding of dynamic environments, necessitating the simultaneous estimation of metric depth, semantic segmentation, and instance trajectories. While depth-aware video panoptic segmentation (DVPS) unifies these tasks, existing approaches often rely on computationally expensive, multi-stage pipelines or offline tracking, rendering them unsuitable for real-time decision-making. To address this, we propose DVPSFormer, a unified online architecture designed for efficient 4D scene understanding. Central to our approach is explicit scene discretization (ESD), a novel mechanism that leverages segmentation queries to represent foreground and background regions, enabling a discrete-to-continuous (D2C) depth head to decode metric depth in a single pass. This tightly couples semantic and geometric learning while significantly reducing latency. Furthermore, we propose an online majority voting (OMV) mechanism that exploits temporal consistency to refine classification during instance tracking. DVPSFormer establishes a new state-of-the-art on the Cityscapes-DVPS and SemKITTI-DVPS benchmarks, offering a streamlined solution for online robotic perception.

• We re-frames panoptic segmentation as explicit scene discretization (ESD) with a discrete-to-continuous depth head for single-pass metric depth, and couples it with lightweight query-based tracking and online majority voting (OMV).
• We target online depth-aware video panoptic segmentation for autonomous driving, addressing the latency and offline-tracking limitations that make existing DVPS pipelines impractical for real-time deployment.
• We achieve state-of-the-art performance on Cityscapes-DVPS and SemKITTI-DVPS while running faster than prior methods.