Automated Quality Check of Sensor Data Annotations
By: Niklas Freund, Zekiye Ilknur-Öz, Tobias Klockau, Patrick Naumann, Philipp Neumaier, Martin Köppel
Published: 2026-03-15
View on arXiv →#cs.CV
Abstract
This paper proposes an automated method for checking the quality of sensor data annotations, a critical component for training reliable machine learning models in autonomous systems. Ensuring high-quality annotations is vital for the safety and performance of AI applications in self-driving cars, robotics, and surveillance.