Semi-Supervised Diversity-Aware Domain Adaptation for 3D Object detection
This addresses domain adaptation for autonomous vehicles, enabling better generalization across regions with minimal annotation effort, though it appears incremental as it builds on existing domain adaptation and continual learning techniques.
The paper tackles the problem of 3D object detectors struggling to generalize across different geographic domains in autonomous driving by proposing a lidar domain adaptation method based on neuron activation patterns, which achieves state-of-the-art performance with only a small, diverse subset of annotated target domain samples.
3D object detectors are fundamental components of perception systems in autonomous vehicles. While these detectors achieve remarkable performance on standard autonomous driving benchmarks, they often struggle to generalize across different domains - for instance, a model trained in the U.S. may perform poorly in regions like Asia or Europe. This paper presents a novel lidar domain adaptation method based on neuron activation patterns, demonstrating that state-of-the-art performance can be achieved by annotating only a small, representative, and diverse subset of samples from the target domain if they are correctly selected. The proposed approach requires very small annotation budget and, when combined with post-training techniques inspired by continual learning prevent weight drift from the original model. Empirical evaluation shows that the proposed domain adaptation approach outperforms both linear probing and state-of-the-art domain adaptation techniques.