GAPrompt: Geometry-Aware Point Cloud Prompt for 3D Vision Model

TL;DR

GAPrompt introduces a geometry-aware prompting method for 3D vision models, significantly improving performance on point cloud tasks with minimal trainable parameters by explicitly incorporating geometric cues.

Contribution

The paper proposes a novel geometry-aware prompting framework that enhances 3D model fine-tuning efficiency by explicitly leveraging geometric information in point clouds.

Findings

Outperforms state-of-the-art PEFT methods on benchmarks

Achieves competitive results with full fine-tuning using only 2.19% trainable parameters

Effectively captures fine-grained and global geometric features

Abstract

Pre-trained 3D vision models have gained significant attention for their promising performance on point cloud data. However, fully fine-tuning these models for downstream tasks is computationally expensive and storage-intensive. Existing parameter-efficient fine-tuning (PEFT) approaches, which focus primarily on input token prompting, struggle to achieve competitive performance due to their limited ability to capture the geometric information inherent in point clouds. To address this challenge, we propose a novel Geometry-Aware Point Cloud Prompt (GAPrompt) that leverages geometric cues to enhance the adaptability of 3D vision models. First, we introduce a Point Prompt that serves as an auxiliary input alongside the original point cloud, explicitly guiding the model to capture fine-grained geometric details. Additionally, we present a Point Shift Prompter designed to extract global shape information from the point cloud, enabling instance-specific geometric adjustments at the input level. Moreover, our proposed Prompt Propagation mechanism incorporates the shape information into the model's feature extraction process, further strengthening its ability to capture essential geometric characteristics. Extensive experiments demonstrate that GAPrompt significantly outperforms state-of-the-art PEFT methods and achieves competitive results compared to full fine-tuning on various benchmarks, while utilizing only 2.19% of trainable parameters. Our code is available at https://github.com/zhoujiahuan1991/ICML2025-GAPrompt.