点聚焦注意力结合上下文扫描状态空间：面向点云表示的鲁棒生物视觉感知
Point-Focused Attention Meets Context-Scan State Space: Robust Biological Visual Perception for Point Cloud Representation
查看机构详情
- College of Artificial Intelligence, Nanjing University of Aeronautics and Astronautics（南京航空航天大学人工智能学院）
- School of Mathematical Sciences, Beijing University of Posts and Telecommunications（北京邮电大学数学科学学院）
机构由 AI 辅助整理，请以论文原文为准。
浏览论文内容
中文总结 AI 辅助
本文提出仿生点云表示学习网络PointLearner，结合点聚焦注意力与上下文扫描状态空间，在多点云任务中实现最优性能且鲁棒性卓越。
中文摘要 AI 辅助
协同捕捉复杂局部结构与全局上下文依赖已成为点云表示学习的关键挑战。为解决该问题，本文提出PointLearner，一种紧密契合生物视觉的点云表示学习网络，采用受中央凹启发的主动处理策略，可同时实现局部几何建模与长程依赖交互。具体而言，首先设计点聚焦注意力，通过局部邻域与空间下采样特征间的竞争归一化注意力机制，模拟视觉焦点处的中央凹视觉；空间下采样特征由基于可学习诱导点的池化方法提取，该方法可通过控制诱导点数量并使其与点云直接交互，灵活适配点云的非均匀分布。其次，提出上下文扫描状态空间，模拟眼睛的扫视推理，通过希尔伯特曲线引导的扫描路径对双向S6进行推理，以获取场景的整体语义结构与空间内容。凭借这种“先聚焦后上下文”的仿生设计，PointLearner展现出卓越的鲁棒性，并在多个点云任务中达到了当前最优性能。
英文摘要
Synergistically capturing intricate local structures and global contextual dependencies has become a critical challenge in point cloud representation learning. To address this, we introduce PointLearner, a point cloud representation learning network that closely aligns with biological vision which employs an active, foveation-inspired processing strategy, thus enabling local geometric modeling and long-range dependency interactions simultaneously. Specifically, we first design a point-focused attention, which simulates foveal vision at the visual focus through a competitive normalized attention mechanism between local neighbors and spatially downsampled features. The spatially downsampled features are extracted by a pooling method based on learnable inducing points, which can flexibly adapt to the non-uniform distribution of point clouds as the number of inducing points is controlled and they interact directly with point clouds. Second, we propose a context-scan state space that mimics eye's saccade inference, which infers the overall semantic structure and spatial content in the scene through a scan path guided by the Hilbert curve for the bidirectional S6. With this focus-then-context biomimetic design, PointLearner demonstrates remarkable robustness and achieves state-of-the-art performance across multiple point cloud tasks.
