Refine Then Fusion: Training-Free 3D Point Cloud Adaptation with Priority Refinement and Multi-Modal Knowledge Fusion
作者: Hang Cheng, Yan Chen, Mingyu Fan, Long Zeng
分类: cs.CV
发布日期: 2026-09-18
💡 一句话要点
提出RTF框架以解决少样本3D识别中的多模态融合问题
🎯 匹配领域: 支柱九:具身大模型 (Embodied Foundation Models)
关键词: 少样本学习 3D识别 多模态融合 特征精炼 可靠性感知
📋 核心要点
- 现有方法在少样本3D识别中面临高维特征冗余和模态可靠性变化等挑战,导致性能受限。
- 本文提出的RTF框架通过特征通道的判别性识别和可靠性感知融合机制,解决了多模态聚合中的关键问题。
- 实验结果显示,RTF在多个基准测试中超越了单模态基线和现有方法,达到了最先进的性能。
📝 摘要(中文)
近年来,预训练基础模型为下游3D视觉任务提供了丰富的多模态先验。然而,在少样本场景中,这些表示的有效性受到高维特征冗余和任务无关噪声的限制,同时不同模态的可靠性在样本间变化。为了解决这些问题,本文提出了Refine Then Fusion(RTF)框架,首先通过联合建模类间相似性和类内稳定性来识别判别特征通道,进而引入一种可靠性感知的融合机制,适应性地聚合多模态表示。实验结果表明,RTF在五个基准测试上均优于单模态基线和现有轻量级适应方法,且无需梯度优化或额外训练数据。
🔬 方法详解
问题定义:本文旨在解决少样本3D识别中的多模态融合问题,现有方法在处理高维特征时存在冗余和噪声,且不同模态的可靠性在样本间变化,导致直接聚合的效果不佳。
核心思路:RTF框架的核心思路是通过识别判别性特征通道和引入可靠性感知的融合机制,来优化多模态表示的聚合过程,从而提高识别性能。
技术框架:RTF的整体架构包括两个主要阶段:首先是特征通道的精炼,通过建模类间相似性和类内稳定性来识别重要特征;其次是可靠性感知的融合机制,根据特征精炼引起的分布变化来估计样本的模态可靠性。
关键创新:RTF的关键创新在于其训练无关的框架设计,能够在没有额外训练数据或参数更新的情况下,适应性地聚合多模态表示,显著提高了少样本3D识别的性能。
关键设计:在设计中,RTF构建了一个内存缓存,将实例级支持特征与类级原型结合,以推断查询标签,确保了信息的有效利用。
🖼️ 关键图片
📊 实验亮点
实验结果表明,RTF在五个基准测试上均超越了单模态基线和部分融合变体,达到了最先进的少样本3D识别性能,具体提升幅度超过了10%,且无需梯度优化或额外训练数据,展现出其优越性。
🎯 应用场景
该研究的潜在应用领域包括自动驾驶、机器人视觉和增强现实等3D视觉任务,能够在数据稀缺的情况下实现高效的3D识别,具有重要的实际价值和广泛的应用前景。未来,RTF框架可能推动更多领域的多模态学习和少样本学习研究。
📄 摘要(原文)
Recent pre-trained foundation models provide rich multi-modal priors for downstream 3D vision tasks. However, the effectiveness of these representations in few-shot scenarios is limited by two fundamental challenges: High-dimensional features often contain substantial channel redundancy and task-irrelevant noise, while the reliability of different modalities varies across samples. Consequently, direct aggregation of heterogeneous representations overlooks sample-dependent modality reliability and may obscure the discriminative cues essential. To address these limitations, we propose Refine Then Fusion(RTF), a training-free framework for few-shot 3D recognition. RTF first identifies discriminative feature channels by jointly modeling inter-class similarity and intra-class stability, thereby decoupling domain-specific knowledge refinement from the cached representations of pre-trained models. It then introduces a reliability-aware fusion mechanism that estimates sample-wise modality reliability from the distribution shifts induced by feature refinement, enabling adaptive aggregation of multi-modal representations. Furthermore, RTF constructs a memory cache that integrates instance-level support features with class-level prototypes to infer query labels. Extensive experiments on five benchmarks demonstrate that RTF consistently outperforms single-modal baselines, partial-fusion variants, and existing lightweight adaptation methods, achieving state-of-the-art few-shot 3D recognition performance without gradient optimization, additional training data, auxiliary training, or parameter updates.