Self-supervised adversarial masking for 3D point cloud representation learning
This work addresses the need for better representation learning in 3D point cloud data, which is incremental as it builds on existing self-supervised methods by optimizing mask selection.
The paper tackles the problem of improving self-supervised representation learning for 3D point clouds by introducing PointCAM, an adversarial method that learns a masking function instead of using random masking, achieving state-of-the-art or competitive performance on various downstream tasks.
Self-supervised methods have been proven effective for learning deep representations of 3D point cloud data. Although recent methods in this domain often rely on random masking of inputs, the results of this approach can be improved. We introduce PointCAM, a novel adversarial method for learning a masking function for point clouds. Our model utilizes a self-distillation framework with an online tokenizer for 3D point clouds. Compared to previous techniques that optimize patch-level and object-level objectives, we postulate applying an auxiliary network that learns how to select masks instead of choosing them randomly. Our results show that the learned masking function achieves state-of-the-art or competitive performance on various downstream tasks. The source code is available at https://github.com/szacho/pointcam.