TeFF: Tracking-enhanced Forgetting-free Few-shot 3D LiDAR Semantic Segmentation
This work addresses the challenge of segmenting newly emerged, unannotated objects in autonomous driving, representing an incremental advance in few-shot 3D LiDAR semantic segmentation.
The paper tackles the few-shot learning problem for 3D LiDAR semantic segmentation in autonomous driving by using tracking to generate pseudo-ground-truths from temporal data, which enhances learning on novel classes but causes catastrophic forgetting, mitigated by incorporating LoRA to preserve base class performance while improving adaptability.
In autonomous driving, 3D LiDAR plays a crucial role in understanding the vehicle's surroundings. However, the newly emerged, unannotated objects presents few-shot learning problem for semantic segmentation. This paper addresses the limitations of current few-shot semantic segmentation by exploiting the temporal continuity of LiDAR data. Employing a tracking model to generate pseudo-ground-truths from a sequence of LiDAR frames, our method significantly augments the dataset, enhancing the model's ability to learn on novel classes. However, this approach introduces a data imbalance biased to novel data that presents a new challenge of catastrophic forgetting. To mitigate this, we incorporate LoRA, a technique that reduces the number of trainable parameters, thereby preserving the model's performance on base classes while improving its adaptability to novel classes. This work represents a significant step forward in few-shot 3D LiDAR semantic segmentation for autonomous driving. Our code is available at https://github.com/junbao-zhou/Track-no-forgetting.