Scholarlink

The scale-up of autonomous vehicles depends heavily on their ability to deal with rare objects on the road. In order to handle such situations, it is necessary to detect anomalies in the first place. Anomaly detection has made great progress in the past years but suffers from poorly designed benchmarks with a strong focus on camera data. In this work, we present AnoVox, the largest benchmark for ANOmaly detection in autonomous driving to date. AnoVox incorporates multimodal sensor data and spatial VOXel ground truth, allowing for the comparison of methods independent of their used sensor. We propose a formal definition of normality and provide a compliant training dataset. AnoVox is the first benchmark to contain both content and temporal anomalies.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

AnoVox: A Benchmark for Multimodal Anomaly Detection in Autonomous Driving

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Daniel Bogdoll,
  • Iramm Hamdard,
  • Lukas Namgyu Rößler,
  • Felix Geisler,
  • Muhammed Bayram,
  • Felix Wang,
  • Jan Imhof,
  • Miguel de Campos,
  • Anushervon Tabarov,
  • Yitian Yang,
  • Martin Gontscharow,
  • Hanno Gottschalk,
  • J. Marius Zöllner

摘要

The scale-up of autonomous vehicles depends heavily on their ability to deal with rare objects on the road. In order to handle such situations, it is necessary to detect anomalies in the first place. Anomaly detection has made great progress in the past years but suffers from poorly designed benchmarks with a strong focus on camera data. In this work, we present AnoVox, the largest benchmark for ANOmaly detection in autonomous driving to date. AnoVox incorporates multimodal sensor data and spatial VOXel ground truth, allowing for the comparison of methods independent of their used sensor. We propose a formal definition of normality and provide a compliant training dataset. AnoVox is the first benchmark to contain both content and temporal anomalies.