VGER: Voxel-Guided Global Event Ranking for Event Cloud Attribution

2026/08/02 by Youxin Jiang, Baoheng Fu, Hongwei Ren +1
Computer Science · #cs.CV #cs.AI

paper · pdf

arxiv created 2026/08/02 · arxiv updated 2026/08/04

Abstract

Event cameras produce sparse and asynchronous event streams that provide rich spatio-temporal information for efficient perception. Recent advances in event-based models have demonstrated strong performance by directly modeling asynchronous events without dense frame reconstruction. However, identifying the event-level evidence behind their predictions is crucial for improving model transparency and reliability. Directly adapting point-level saliency methods from point clouds provides fine-grained attribution but overlooks event-specific spatio-temporal structures. To address this limitation, we propose Voxel-Guided Global Event Ranking (VGER), a training-free attribution framework for point-based event cloud networks. VGER combines event-level gradient evidence with task-aware voxel perturbation evidence, transferring regional contribution into event-level attribution scores while preserving fine-grained resolution. Furthermore, VGER introduces a unified event ranking strategy, where high-ranked events are expected to be prediction-critical and low-ranked events are expected to have limited influence on predictions. We evaluate VGER on three event-based benchmarks with PointNet, PointNet++, and EventMamba. Across nine dataset-backbone settings, VGER consistently improves both high-tail and low-tail deletion performance over point-level saliency baselines.

Citations

• XAI for Point Cloud Data using Perturbations based on Meaningful Segmentation
• Scalable Event Cloud Network for Event-based Classification
• FFAM: Feature Factorization Activation Map for Explanation of 3D Detectors
• A Simple and Effective Point-based Network for Event Camera 6-DOFs Pose Relocalization
• Fast and Simple Explainability for Point Cloud Networks
• E2PNet: Event to Point Cloud Registration with Spatio-Temporal Representation Learning
• SpikePoint: An Efficient Point-based Spiking Neural Network for Event Cameras Action Recognition
• GET: Group Event Transformer for Event-Based Vision
• Event Transformer. A sparse-aware solution for efficient event data processing
• Metrics for saliency map evaluation of deep learning explanation methods
• EventPoint: Self-Supervised Interest Point Detection and Description for Event-based Camera
• A Voxel Graph CNN for Object Classification with Event Cameras
• Sanity Checks for Saliency Metrics
• End-to-End Learning of Representations for Asynchronous Event-Based Data
• On the (In)fidelity and Sensitivity for Explanations
• EventNet: Asynchronous Recursive Event Processing
• PointCloud Saliency Maps
• A Benchmark for Interpretability Methods in Deep Neural Networks
• On the Robustness of Interpretability Methods
• RISE: Randomized Input Sampling for Explanation of Black-box Models
• HATS: Histograms of Averaged Time Surfaces for Robust Event-based Object\n Classification
• Dynamic Graph CNN for Learning on Point Clouds
• Dynamic Graph CNN for Learning on Point Clouds
• The (Un)reliability of saliency methods
• SmoothGrad: removing noise by adding noise
• PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space
• A Unified Approach to Interpreting Model Predictions
• PointNet: Deep Learning on Point Sets for 3D Classification and\n Segmentation
• Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization
• "Why Should I Trust You?": Explaining the Predictions of Any Classifier
• Converting Static Image Datasets to Spiking Neuromorphic Datasets Using Saccades
• Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps
• Visualizing and Understanding Convolutional Networks