Pre-training Point Cloud Compact Model with Partial-aware Reconstruction

TL;DR

This paper introduces Point-CPR, a compact pre-training method for point cloud models that improves robustness and efficiency by partial-aware reconstruction, outperforming larger models on various tasks.

Contribution

The authors propose a novel partial-aware reconstruction approach and a compact encoder, significantly reducing model size and enhancing robustness compared to existing Masked Point Modeling methods.

Findings

Outperforms state-of-the-art MPM models with only 2% of parameters

Enhances robustness by preventing coordinate leakage in the decoder

Reduces computational requirements with a compact encoder

Abstract

The pre-trained point cloud model based on Masked Point Modeling (MPM) has exhibited substantial improvements across various tasks. However, two drawbacks hinder their practical application. Firstly, the positional embedding of masked patches in the decoder results in the leakage of their central coordinates, leading to limited 3D representations. Secondly, the excessive model size of existing MPM methods results in higher demands for devices. To address these, we propose to pre-train Point cloud Compact Model with Partial-aware \textbf{R}econstruction, named Point-CPR. Specifically, in the decoder, we couple the vanilla masked tokens with their positional embeddings as randomly masked queries and introduce a partial-aware prediction module before each decoder layer to predict them from the unmasked partial. It prevents the decoder from creating a shortcut between the central coordinates of masked patches and their reconstructed coordinates, enhancing the robustness of models. We also devise a compact encoder composed of local aggregation and MLPs, reducing the parameters and computational requirements compared to existing Transformer-based encoders. Extensive experiments demonstrate that our model exhibits strong performance across various tasks, especially surpassing the leading MPM-based model PointGPT-B with only 2% of its parameters.