Masked point modeling has become a promising scheme of self-supervised pre-training for point clouds. Existing methods reconstruct either the original points or related features as the objective of pre-training. However, considering the diversity of downstream tasks, it is necessary for the model to have both low- and high-level representation modeling capabilities to capture geometric details and semantic contexts during pre-training. To this end, M3CS is proposed to enable the model with the above abilities. Specifically, with masked point cloud as input, M3CS introduces two decoders to predict masked representations and the original points simultaneously. While an extra decoder doubles parameters for the decoding process and may lead to overfitting, we propose siamese decoders to keep the amount of learnable parameters unchanged. Further, we propose an online codebook projecting continuous tokens into discrete ones before reconstructing masked points. In such way, we can enforce the decoder to take effect through the combinations of tokens rather than remembering each token. Comprehensive experiments show that M3CS achieves superior performance at both classification and segmentation tasks, outperforming existing methods.
@article{arxiv.2309.13235,
title = {M$^3$CS: Multi-Target Masked Point Modeling with Learnable Codebook and Siamese Decoders},
author = {Qibo Qiu and Honghui Yang and Wenxiao Wang and Shun Zhang and Haiming Gao and Haochao Ying and Wei Hua and Xiaofei He},
journal= {arXiv preprint arXiv:2309.13235},
year = {2023}
}