Uncertainty-Aware Vision-based Risk Object Identification via Conformal Risk Tube Prediction

2026-03-25

AI Summary

This work addresses the limitations of existing vision-based risk object recognition methods, which rely on deterministic decisions and neglect uncertainty, often leading to misjudgments, temporal instability, and safety-critical failures in complex multi-risk interaction scenarios. To overcome these challenges, we propose the Conformal Risk Tube Prediction framework—a principled approach that unifies the modeling of spatiotemporal risk uncertainty and introduces a conformal prediction mechanism with guaranteed risk coverage, yielding well-calibrated risk scores. Our contributions include the novel framework itself, a new simulation dataset supporting coupled multi-risk interactions, and tailored evaluation metrics. Experimental results demonstrate that our method significantly enhances recognition robustness, effectively reduces false-positive emergency braking alerts in autonomous driving, and improves downstream task performance.

Technology Category

• Computer Vision: Multi-modal Vision
• Reasoning under Uncertainty: Uncertainty Representations
• Machine Learning: Calibration & Uncertainty Quantification

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

We study object importance-based vision risk object identification (Vision-ROI), a key capability for hazard detection in intelligent driving systems. Existing approaches make deterministic decisions and ignore uncertainty, which could lead to safety-critical failures. Specifically, in ambiguous scenarios, fixed decision thresholds may cause premature or delayed risk detection and temporally unstable predictions, especially in complex scenes with multiple interacting risks. Despite these challenges, current methods lack a principled framework to model risk uncertainty jointly across space and time. We propose Conformal Risk Tube Prediction, a unified formulation that captures spatiotemporal risk uncertainty, provides coverage guarantees for true risks, and produces calibrated risk scores with uncertainty estimates. To conduct a systematic evaluation, we present a new dataset and metrics probing diverse scenario configurations with multi-risk coupling effects, which are not supported by existing datasets. We systematically analyze factors affecting uncertainty estimation, including scenario variations, per-risk category behavior, and perception error propagation. Our method delivers substantial improvements over prior approaches, enhancing vision-ROI robustness and downstream performance, such as reducing nuisance braking alerts. For more qualitative results, please visit our project webpage: https://hcis-lab.github.io/CRTP/

Problem

• risk object identification
• uncertainty quantification
• spatiotemporal uncertainty
• vision-based perception
• intelligent driving

Innovation

• Conformal Risk Tube Prediction
• Uncertainty-Aware Vision
• Spatiotemporal Risk Modeling
• Risk Object Identification
• Calibrated Risk Scores