Reis, Philipp (Research Center for Information Technology), Ransiek, Joshua (FZI Research Center for Information Technology), Langner, Jacob (FZI Research Center for Information Technology), Sax, Eric (Karlsruhe Institute of Technology), Petri, David (Karlsruhe Institute of Technology), Schürmann, Tobias (Daimler AG)

A Data-Driven Novelty Score for Diverse In-Vehicle Data Recording

Scheduled for presentation during the Regular Session "S07b-Smart Infrastructure and Data-Driven Sensing for Intelligent Mobility" (WE-EA-T7), Wednesday, November 19, 2025, 13:50−14:10, Coolangata 1

2025 IEEE 28th International Conference on Intelligent Transportation Systems (ITSC), November 18-21, 2025, Gold Coast, Australia

This information is tentative and subject to change. Compiled on October 19, 2025

Abstract

High-quality datasets are essential for training robust perception systems in autonomous driving. However, real-world data collection is often biased toward common scenes and objects, leaving novel cases underrepresented. This imbalance hinders model generalization and compromises safety. The core issue is the curse of rarity. Over time, novel events occur infrequently, and standard logging methods fail to capture them effectively. As a result, large volumes of redundant data are stored, while critical novel cases are diluted, leading to biased datasets. This work presents a real-time data selection method focused on object-level novelty detection to build more balanced and diverse datasets. The method assigns a data-driven novelty score to image frames using a novel adaptive Mean Shift algorithm. It models normal content based on mean and covariance statistics to identify frames with novel objects, discarding those with redundant elements. The main findings show that reducing the training dataset size with this method can improve model performance, whereas higher redundancy tends to degrade it. Moreover, as data redundancy increases, more aggressive filtering becomes both possible and beneficial. While random sampling can offer some gains, it often leads to overfitting and unpredictability in outcomes. The proposed method supports real-time deployment with 32 frames per second and is constant over time. By continuously updating the definition of normal content, it enables efficient detection of novelties in a continuous data stream.