Learning to Identify Out-of-Distribution Objects for 3D LiDAR Anomaly Segmentation
For autonomous driving and robotics, this work addresses the limited research in 3D anomaly segmentation by providing an efficient feature-space method and new datasets, though the novelty is incremental as it adapts 2D techniques to 3D.
The paper proposes a new method for 3D LiDAR anomaly segmentation that models feature distributions of inlier classes to detect out-of-distribution objects, and introduces mixed real-synthetic datasets to address domain gaps. The approach achieves state-of-the-art results on existing real-world datasets and competitive results on the new datasets.
Understanding the surrounding environment is fundamental in autonomous driving and robotic perception. Distinguishing between known classes and previously unseen objects is crucial in real-world environments, as done in Anomaly Segmentation. However, research in the 3D field remains limited, with most existing approaches applying post-processing techniques from 2D vision. To cover this lack, we propose a new efficient approach that directly operates in the feature space, modeling the feature distribution of inlier classes to constrain anomalous samples. Moreover, the only publicly available 3D LiDAR anomaly segmentation dataset contains simple scenarios, with few anomaly instances, and exhibits a severe domain gap due to its sensor resolution. To bridge this gap, we introduce a set of mixed real-synthetic datasets for 3D LiDAR anomaly segmentation, built upon established semantic segmentation benchmarks, with multiple out-of-distribution objects and diverse, complex environments. Extensive experiments demonstrate that our approach achieves state-of-the-art and competitive results on the existing real-world dataset and the newly introduced mixed datasets, respectively, validating the effectiveness of our method and the utility of the proposed datasets. Code and datasets are available at https://simom0.github.io/lido-page/.