Stereo-based 3D Anomaly Object Detection for Autonomous Driving: A New Dataset and Baseline

TL;DR

This paper introduces S3AD, a stereo-based 3D anomaly detection algorithm for autonomous driving, along with a new dataset KITTI-AR that enhances generalization and zero-shot detection of rare anomalies.

Contribution

It proposes a decoupled training strategy for 3D and 2D detection and creates a new augmented reality stereo dataset with diverse categories for better anomaly detection.

Findings

S3AD achieves improved anomaly detection performance.

KITTI-AR dataset extends detection capabilities to 97 new categories.

Zero-shot detection is effectively evaluated on KITTI-AR-OoD subset.

Abstract

3D detection technology is widely used in the field of autonomous driving, with its application scenarios gradually expanding from enclosed highways to open conventional roads. For rare anomaly categories that appear on the road, 3D detection models trained on closed sets often misdetect or fail to detect anomaly objects. To address this risk, it is necessary to enhance the generalization ability of 3D detection models for targets of arbitrary shapes and to possess the capability to filter out anomalies. The generalization of 3D detection is limited by two factors: the coupled training of 2D and 3D, and the insufficient diversity in the scale distribution of training samples. This paper proposes a Stereo-based 3D Anomaly object Detection (S3AD) algorithm, which decouples the training strategy of 3D and 2D to release the generalization ability for arbitrary 3D foreground detection, and proposes an anomaly scoring algorithm based on foreground confidence prediction, achieving target-level anomaly scoring. In order to further verify and enhance the generalization of anomaly detection, we use a 3D rendering method to synthesize two augmented reality binocular stereo 3D detection datasets which named KITTI-AR. KITTI-AR extends upon KITTI by adding 97 new categories, totaling 6k pairs of stereo images. The KITTI-AR-ExD subset includes 39 common categories as extra training data to address the sparse sample distribution issue. Additionally, 58 rare categories form the KITTI-AR-OoD subset, which are not used in training to simulate zero-shot scenarios in real-world settings, solely for evaluating 3D anomaly detection. Finally, the performance of the algorithm and the dataset is verified in the experiments. (Code and dataset can be obtained at https://github.com/shiyi-mu/S3AD-Code).