Learning Efficient Meshflow and Optical Flow from Event Cameras

2025-10-05

AI Summary

This work addresses the lack of dedicated datasets and robust methods for meshflow estimation from event cameras—particularly under varying event densities. To this end, we introduce HREM, the first large-scale, high-resolution meshflow dataset, and propose EEMFlow, a lightweight network featuring a customized encoder-decoder architecture. EEMFlow integrates a confidence-guided detail completion (CDC) module and an adaptive density module (ADM) to enable spatially smooth motion field prediction across multi-density event inputs. Experiments demonstrate that EEMFlow achieves a 30× speedup over existing state-of-the-art methods. Incorporating ADM yields 8–10% accuracy gains on both sparse and dense scenes. Moreover, in joint evaluation of meshflow and optical flow, EEMFlow attains the optimal trade-off between accuracy and efficiency.

Technology Category

• Computer Vision: Motion & Tracking
• Intelligent Robots: State Estimation
• Machine Learning: Hardware-aware ML

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Data management and stream processing for Web, mobile and wireless applications
• Security and Privacy: Large-scale security measurements

Abstract

In this paper, we explore the problem of event-based meshflow estimation, a novel task that involves predicting a spatially smooth sparse motion field from event cameras. To start, we review the state-of-the-art in event-based flow estimation, highlighting two key areas for further research: i) the lack of meshflow-specific event datasets and methods, and ii) the underexplored challenge of event data density. First, we generate a large-scale High-Resolution Event Meshflow (HREM) dataset, which showcases its superiority by encompassing the merits of high resolution at 1280x720, handling dynamic objects and complex motion patterns, and offering both optical flow and meshflow labels. These aspects have not been fully explored in previous works. Besides, we propose Efficient Event-based MeshFlow (EEMFlow) network, a lightweight model featuring a specially crafted encoder-decoder architecture to facilitate swift and accurate meshflow estimation. Furthermore, we upgrade EEMFlow network to support dense event optical flow, in which a Confidence-induced Detail Completion (CDC) module is proposed to preserve sharp motion boundaries. We conduct comprehensive experiments to show the exceptional performance and runtime efficiency (30x faster) of our EEMFlow model compared to the recent state-of-the-art flow method. As an extension, we expand HREM into HREM+, a multi-density event dataset contributing to a thorough study of the robustness of existing methods across data with varying densities, and propose an Adaptive Density Module (ADM) to adjust the density of input event data to a more optimal range, enhancing the model's generalization ability. We empirically demonstrate that ADM helps to significantly improve the performance of EEMFlow and EEMFlow+ by 8% and 10%, respectively. Code and dataset are released at https://github.com/boomluo02/EEMFlowPlus.

Problem

• Estimating sparse motion fields from event camera data
• Addressing the lack of event-based meshflow datasets and methods
• Handling varying event data densities for robust flow estimation

Innovation

• Created high-resolution event meshflow dataset HREM
• Proposed lightweight EEMFlow network for meshflow estimation
• Introduced adaptive density module to enhance generalization