Scholarlink

<p>From a perspective of feature matching, optical flow estimation for event cameras involves identifying event correspondences by comparing feature similarity between two consecutive event slices. In this work, we introduce an effective and robust high-dimensional feature descriptor for event slices, utilizing Vector Symbolic Architectures (VSA). The topological similarity among neighboring variables within VSA contributes to the enhanced representation similarity of feature descriptors for flow-matching points, while its structured symbolic representation capacity facilitates feature fusion from both event polarities and multiple spatial scales. Based on this high-dimensional feature descriptor, we propose a novel feature matching framework for event-based optical flow, encompassing both model-based (VSA-Flow) and self-supervised learning (VSA-SM) methods. In VSA-Flow, accurate optical flow estimation validates the effectiveness of feature descriptors. In VSA-SM, a novel similarity maximization method based on the feature descriptor is proposed to learn optical flow in a self-supervised way from events alone, eliminating the need for auxiliary grayscale images. Evaluation results demonstrate that our VSA-based method is highly effective compared to both model-based and self-supervised learning methods on the DSEC and MVSEC benchmarks. This contribution marks a significant advancement in event-based optical flow within the feature matching methodology. Our code will be available at <a href="https://github.com/youhongzhi/VSA_For_EventOpticalFlow">https://github.com/youhongzhi/VSA_For_EventOpticalFlow</a>.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Vector-Symbolic Architecture for Event-Based Optical Flow

  • 期刊论文

  • 发表日期：

    2026年7月27日

查看全文
  • Hongzhi You,
  • Yijun Cao,
  • Wei Yuan,
  • Fanjun Wang,
  • Ning Qiao,
  • Yongjie Li

摘要

From a perspective of feature matching, optical flow estimation for event cameras involves identifying event correspondences by comparing feature similarity between two consecutive event slices. In this work, we introduce an effective and robust high-dimensional feature descriptor for event slices, utilizing Vector Symbolic Architectures (VSA). The topological similarity among neighboring variables within VSA contributes to the enhanced representation similarity of feature descriptors for flow-matching points, while its structured symbolic representation capacity facilitates feature fusion from both event polarities and multiple spatial scales. Based on this high-dimensional feature descriptor, we propose a novel feature matching framework for event-based optical flow, encompassing both model-based (VSA-Flow) and self-supervised learning (VSA-SM) methods. In VSA-Flow, accurate optical flow estimation validates the effectiveness of feature descriptors. In VSA-SM, a novel similarity maximization method based on the feature descriptor is proposed to learn optical flow in a self-supervised way from events alone, eliminating the need for auxiliary grayscale images. Evaluation results demonstrate that our VSA-based method is highly effective compared to both model-based and self-supervised learning methods on the DSEC and MVSEC benchmarks. This contribution marks a significant advancement in event-based optical flow within the feature matching methodology. Our code will be available at https://github.com/youhongzhi/VSA_For_EventOpticalFlow.