CoSMix: Compositional Semantic Mix for Domain Adaptation in 3D LiDAR Segmentation
This addresses the problem of model generalization across different sensors and environments for autonomous driving, representing an incremental advance by applying sample mixing to point cloud data.
The paper tackles domain adaptation for 3D LiDAR segmentation by proposing CoSMix, a sample mixing method that processes synthetic and real-world point clouds, and it outperforms state-of-the-art methods by a large margin on two large-scale datasets.
3D LiDAR semantic segmentation is fundamental for autonomous driving. Several Unsupervised Domain Adaptation (UDA) methods for point cloud data have been recently proposed to improve model generalization for different sensors and environments. Researchers working on UDA problems in the image domain have shown that sample mixing can mitigate domain shift. We propose a new approach of sample mixing for point cloud UDA, namely Compositional Semantic Mix (CoSMix), the first UDA approach for point cloud segmentation based on sample mixing. CoSMix consists of a two-branch symmetric network that can process labelled synthetic data (source) and real-world unlabelled point clouds (target) concurrently. Each branch operates on one domain by mixing selected pieces of data from the other one, and by using the semantic information derived from source labels and target pseudo-labels. We evaluate CoSMix on two large-scale datasets, showing that it outperforms state-of-the-art methods by a large margin. Our code is available at https://github.com/saltoricristiano/cosmix-uda.