DAP-MAE: Domain-Adaptive Point Cloud Masked Autoencoder for Effective Cross-Domain Learning
It addresses data scarcity in 3D point cloud analysis for researchers and practitioners, offering a domain-adaptive pre-training method that is incremental over existing MAE approaches.
The paper tackles the problem of limited point cloud data across domains for masked autoencoder pre-training by proposing DAP-MAE, which adaptively integrates cross-domain knowledge to improve downstream tasks, achieving 95.18% in object classification and 88.45% in facial expression recognition.
Compared to 2D data, the scale of point cloud data in different domains available for training, is quite limited. Researchers have been trying to combine these data of different domains for masked autoencoder (MAE) pre-training to leverage such a data scarcity issue. However, the prior knowledge learned from mixed domains may not align well with the downstream 3D point cloud analysis tasks, leading to degraded performance. To address such an issue, we propose the Domain-Adaptive Point Cloud Masked Autoencoder (DAP-MAE), an MAE pre-training method, to adaptively integrate the knowledge of cross-domain datasets for general point cloud analysis. In DAP-MAE, we design a heterogeneous domain adapter that utilizes an adaptation mode during pre-training, enabling the model to comprehensively learn information from point clouds across different domains, while employing a fusion mode in the fine-tuning to enhance point cloud features. Meanwhile, DAP-MAE incorporates a domain feature generator to guide the adaptation of point cloud features to various downstream tasks. With only one pre-training, DAP-MAE achieves excellent performance across four different point cloud analysis tasks, reaching 95.18% in object classification on ScanObjectNN and 88.45% in facial expression recognition on Bosphorus.