研究人员开发了HorizonStream，一种新颖的Transformer架构，旨在通过流式视频改进长视域3D重建。该方法通过证据影响核（evidence influence kernel）对几何传播进行建模，解决了内存和时间复杂度问题。HorizonStream通过从短训练片段泛化到超过10,000帧的序列，并保持恒定的内存和线性时间复杂度，实现了最先进的性能。此外，GHOST框架提供了一种无需训练的方法来管理流式3D重建中的键值缓存（key-value cache），将缓存大小减少近一半并加快推理速度。 AI
影响 这些在流式3D重建方面的进展有望在机器人、增强现实和自主系统中实现更强大、更高效的实时应用。
排序理由 多篇论文介绍了3D重建的新方法。
在 Hugging Face Daily Papers 阅读 →
AI 生成摘要 · Google Gemini · 来自 4 个来源。 我们如何撰写摘要 →
HorizonStream addresses long-term 3D reconstruction challenges by modeling geometric propagation through an evidence influence kernel, enabling stable, scalable streaming reconstruction with constant memory and linear time complexity.
arXiv:2605.15852v2 Announce Type: replace Abstract: Streaming 3D reconstruction from long monocular video sequences requires maintaining a key-value (KV) cache that grows linearly with sequence length, creating a severe memory bottleneck. Existing approaches either truncate the c…
arXiv:2605.23889v1 Announce Type: new Abstract: Online 3D reconstruction requires estimating camera pose and scene geometry under strict causal and bounded-memory constraints. Existing methods often suffer from drift, jitter, or collapse on long sequences. We trace these failures…
Online 3D reconstruction requires estimating camera pose and scene geometry under strict causal and bounded-memory constraints. Existing methods often suffer from drift, jitter, or collapse on long sequences. We trace these failures to a fundamental mismatch. Streaming geometry i…