Mask IPL: Noise-Free Intrinsic Position Learning via Computation Graph Clipping for Event-Based Spike-Driven Tracking

📄 arXiv: 2609.18716v1 📥 PDF

作者: Yimeng Shan, Malu Zhang

分类: cs.CV

发布日期: 2026-09-16

💡 一句话要点

提出Mask IPL以解决事件驱动跟踪中的噪声问题

🎯 匹配领域: 支柱八：物理动画 (Physics-based Animation)

关键词: 脉冲神经网络 事件驱动跟踪 内在位置学习 计算图剪裁 多阶段卷积 跟踪精度提升 视觉系统

📋 核心要点

  1. 现有的内在位置学习（IPL）方法在事件驱动跟踪中引入了噪声，导致推理误差和参数收敛问题。
  2. 本文提出了Mask IPL，通过计算图剪裁方法消除噪声，使得前向和反向传播中的无效区域等同于零填充。
  3. 实验结果显示，Mask IPL在不增加参数的情况下，显著提升了Tiny-scale和Base-scale跟踪器的性能。

📝 摘要（中文）

脉冲神经网络（SNN）与事件相机的事件驱动特性相匹配，自然提取时空特征。内在位置学习（IPL）在事件驱动跟踪中获取强位置信息，但其有效性机制缺乏系统理论分析。本文分析表明，IPL在前向和反向传播中引入噪声，导致推理误差和参数收敛问题。我们提出了一种计算图剪裁方法，通过有效性掩码消除噪声，使实际梯度与理想梯度一致，命名为Mask IPL。该方法在不增加参数和计算成本的情况下，显著提升了Tiny-scale跟踪器在多个数据集上的AUC表现。

🔬 方法详解

问题定义：本文旨在解决内在位置学习（IPL）在事件驱动跟踪中引入的噪声问题，导致推理误差和参数收敛不佳。

核心思路：通过提出Mask IPL，利用计算图剪裁方法消除无效区域的噪声，使得前向和反向传播中的梯度更接近理想梯度。

技术框架：整体架构包括多阶段卷积和计算图剪裁，前者提取时空特征，后者通过有效性掩码处理每一层的操作。

关键创新：Mask IPL的核心创新在于通过有效性掩码消除噪声，而不增加额外参数，与传统IPL方法相比，显著提高了跟踪精度。

关键设计：在设计中，采用了零块作为卷积的零填充，确保每次参数更新都由感知模板与搜索帧之间相对位移的梯度驱动。

🖼️ 关键图片

img_0
img_1
img_2

📊 实验亮点

实验结果表明，Mask IPL在Tiny-scale跟踪器上提升了AUC表现，具体在FE108、FELT和VisEvent数据集上均有显著改善，且Base-scale跟踪器的性能也得到了持续提升。

🎯 应用场景

该研究在事件驱动视觉跟踪、机器人导航和自动驾驶等领域具有广泛的应用潜力。通过提高跟踪精度，Mask IPL可以增强系统的稳定性和可靠性，推动智能视觉系统的发展。

📄 摘要（原文）

Spiking Neural Networks (SNNs) match the event-driven nature of event cameras and naturally extract spatiotemporal features. These properties have motivated a series of recent studies on event-based tracking with SNNs. Intrinsic Position Learning (IPL) acquires strong position information without introducing additional parameters, making it a mainstream approach for position encoding in event-based spike-driven tracking. However, the mechanism behind its effectiveness lacks systematic theoretical analysis. Moreover, our analysis reveals that IPL introduces noise in both forward and backward propagation. The former increases inference error, while the latter prevents parameters from converging to better solutions. This paper presents a systematic analysis of IPL and demonstrates that its effectiveness stems from the synergy between IPL and multi-stage convolution. The zero blocks in the joint tensor act as zero padding for convolution, and the resulting boundary effect propagates layer by layer through multi-stage convolution. Every parameter update is therefore driven by a gradient that perceives the relative displacement between template and search frames. Positional encoding added after the convolutional stage cannot provide this information. We further propose a simple Computation Graph Clipping method that applies a validity mask determined by the layout to the operations of every layer, making invalid regions equivalent to zero padding in both forward and backward propagation. This eliminates the noise without introducing additional parameters and makes the actual gradient coincide with the ideal gradient. We name the improved method Mask IPL. Without increasing parameters or computational cost, Mask IPL improves the AUC of the Tiny-scale tracker on FE108, FELT, and VisEvent, and consistently improves the Base-scale tracker as well.