发表机构
Institute of Engineering Sciences, Universidad de O’Higgins; The Iniciativa de Datos e Inteligencia Artificial, University of Chile（奥伊金斯大学工程科学学院; 智利大学数据与人工智能倡议）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究提出eBIRD框架，结合DDPM与ControlNet实现事件引导强度图像重建，在N-MNIST和RGBE-Gaze数据集上验证了通用与专用扩散学习策略的效果，表明可控扩散模型是该任务的有前景方法。
AI 中文摘要
从事件流中重建强度图像仍是一项具有挑战性的问题，因为事件数据具有二值性、稀疏性和异步性。本研究提出eBIRD，一种结合DDPM与基于ControlNet的条件控制的事件引导重建框架。我们针对手写数字（N-MNIST）和人脸（RGBE-Gaze）重建，使用33毫秒事件窗口分析通用和专用的扩散学习策略。在N-MNIST数据集上，通用模型实现最佳重建质量（MSE 0.0052，SSIM 0.8982，PSNR 23.34dB）；而在RGBE-Gaze数据集上，专用模型表现最佳（MSE 0.0161，SSIM 0.7605，PSNR 19.08dB）。这些初步结果表明，可控扩散模型是事件引导强度图像重建的有前景方法，同时凸显最佳学习策略取决于重建领域。
英文摘要
Intensity-image reconstruction from event streams remains a challenging problem due to the binary, sparse, and asynchronous nature of event data. This work proposes eBIRD, an event-guided reconstruction framework that combines a DDPM with ControlNet-based conditioning. We analyze generic and specialized diffusion learning strategies for handwritten digit (N-MNIST) and face (RGBE-Gaze) reconstruction using 33ms event windows. On N-MNIST, the general model achieves the best reconstruction quality (MSE 0.0052, SSIM 0.8982, PSNR 23.34dB), whereas the specialized model performs best on RGBE-Gaze (MSE 0.0161, SSIM 0.7605, PSNR 19.08dB). These preliminary results suggest that controllable diffusion models are a promising approach for event-guided intensity-image reconstruction, while highlighting that the preferred learning strategy depends on the reconstruction domain.
Journal refThe 19th European Conference on Computer Vision Workshops (ECCVW 2026); Workshop on Neuromorphic Vision (NeVi)
