Point-DAE: Denoising Autoencoders for Self-supervised Point Cloud Learning
This work addresses the problem of improving self-supervised learning for point cloud data in 3D vision, offering a more general approach with complementary corruptions, though it is incremental as it builds on masked autoencoder methods.
The paper tackled self-supervised point cloud learning by proposing Point-DAE, a denoising autoencoder that uses various corruptions beyond masking, including affine transformations, to reconstruct original point clouds, achieving effectiveness validated across multiple tasks like object classification and part segmentation.
Masked autoencoder has demonstrated its effectiveness in self-supervised point cloud learning. Considering that masking is a kind of corruption, in this work we explore a more general denoising autoencoder for point cloud learning (Point-DAE) by investigating more types of corruptions beyond masking. Specifically, we degrade the point cloud with certain corruptions as input, and learn an encoder-decoder model to reconstruct the original point cloud from its corrupted version. Three corruption families (\ie, density/masking, noise, and affine transformation) and a total of fourteen corruption types are investigated with traditional non-Transformer encoders. Besides the popular masking corruption, we identify another effective corruption family, \ie, affine transformation. The affine transformation disturbs all points globally, which is complementary to the masking corruption where some local regions are dropped. We also validate the effectiveness of affine transformation corruption with the Transformer backbones, where we decompose the reconstruction of the complete point cloud into the reconstructions of detailed local patches and rough global shape, alleviating the position leakage problem in the reconstruction. Extensive experiments on tasks of object classification, few-shot learning, robustness testing, part segmentation, and 3D object detection validate the effectiveness of the proposed method. The codes are available at \url{https://github.com/YBZh/Point-DAE}.