论文概要
研究领域: CV 作者: Cheng-De Fan, Chun-Wei Tuan Mu, Chen-Wei Chang, Chin-Yang Lin, Kun-Ru Wu, Yu-Chee Tseng, Yu-Lun Liu 发布时间: 2026-07-09 arXiv: 2607.08770
中文摘要
从稀疏事件流中恢复高质量视频是一项具有挑战性的任务。回归方法往往会模糊纹理，而现有生成模型难以保持长期稳定性。我们提出LongE2V，一种利用预训练视频扩散先验来联合处理基于事件的视频重建、预测和帧插值的新方法。通过对基础视频模型进行微调，我们的方法实现了高数据效率和卓越的感知质量。我们引入自回归展开和自适应上下文切换来缓解极长序列中的时间漂移。此外，我们还提出重编码对齐与交叉残差校正，以确保帧插值期间精确的双向一致性。事件体素密度增强则确保了在不同传感器分辨率下的鲁棒性。在真实世界基准上的大量实验表明，LongE2V在所有三项任务上均优于最先进的方法，展现出卓越的时间一致性和零样本泛化能力。项目页面：https://cdfan0627.github.io/LongE2V-page/
原文摘要
Recovering high-quality video from sparse event streams is a challenging task. Regression methods often blur textures, while existing generative models struggle with long-term stability. We propose LongE2V, a novel approach that leverages pre-trained video diffusion priors to jointly handle event-based video reconstruction, prediction, and frame interpolation. By fine-tuning a foundational video model, our approach achieves high data efficiency and superior perceptual quality. We introduce Autoregressive Unrolling and Adaptive Context Switching to mitigate temporal drift in extremely long sequences. We also propose Reencoding Alignment with Cross Residual Correction to ensure precise bidirectional consistency during frame interpolation. Furthermore, Event Voxel Density Augmentation ensures…
— 自动采集于 2026-07-12
#论文 #arXiv #CV #小凯