Hyper-V2X: Hypernetworks for Estimating Epistemic and Aleatoric Uncertainty in Cooperative Bird's-Eye-View Semantic Segmentation

2026-05-20

AI Summary

This work addresses the challenge of effectively quantifying both epistemic and aleatoric uncertainties in existing vehicle-to-everything (V2X) cooperative perception methods. The authors propose a lightweight, architecture-agnostic Bayesian hypernetwork framework that, for the first time, integrates hypernetwork mechanisms into collaborative bird’s-eye-view (BEV) perception. By leveraging V2X contextual embeddings and a partial weight generation strategy, the framework dynamically models the weight distribution of semantic segmentation models to jointly estimate dual uncertainties. Experiments on the OPV2V benchmark demonstrate that the proposed approach significantly enhances perception reliability and yields well-calibrated, accurate uncertainty estimates with negligible additional computational overhead.

Abstract

Cooperative perception enabled by Vehicle-to-Everything (V2X) communication enhances autonomous driving safety by creating a unified environmental representation through shared sensory data. While recent works have advanced multi-agent fusion for improved perception, uncertainty quantification in such cooperative frameworks remains largely unexplored. This paper introduces Hyper-V2X, a hypernetwork-based framework for estimating both epistemic and aleatoric uncertainties in V2X-based perception. Specifically, we propose a partial weight generation scheme and V2X context embedding module that conditions a Bayesian hypernetwork on fused multi-agent features to generate weight distributions for stochastic Bird's-Eye-View (BEV) segmentation. Unlike existing deterministic BEV models, Hyper-V2X enables efficient uncertainty estimation with little computation overhead. Our approach is architecture-agnostic, and can be seamlessly integrating with modern cooperative backbones such as CoBEVT. Experiments on the OPV2V benchmark demonstrate that Hyper-V2X provides accurate, well-calibrated uncertainty estimates and improves overall perception reliability. Our code and benchmark are publicly available under an open-source license: https://github.com/abhishekjagtap1/Hyper-V2X

Problem

• V2X
• uncertainty quantification
• cooperative perception
• Bird's-Eye-View segmentation
• epistemic and aleatoric uncertainty

Innovation

• hypernetworks
• uncertainty quantification
• V2X
• BEV segmentation
• cooperative perception