跳到正文
arXiv cs.AI· Yifan Chang, Jian Yang, Jiangnan Shao, Xiaofeng Wang, Zhaoxin Fan, Renjing Xu, Ming Lu, Xingang Wang, Zheng Zhu·· 4 小时前AI 评分39

MAVE:从电影镜头重建可编辑 3D 布局,配套 MAVEBench 基准与 MAVEData-3K 数据集

MAVE: Reconstructing Editable 3D Blockouts from Film Shots

AI 导读

MAVE 是一个视频转 3D blockout 框架,结合度量观测、智能体建模与基于证据的验证,输出可编辑 Blender 场景和帧对齐的 blockout 视频。

正文 · 原文

View PDF HTML (experimental)

Abstract:Animated 3D blockouts support previsualization and provide structural conditions for video generation, but manually reconstructing them from reference videos is costly. This limits both their use by creators and the production of paired training data. We introduce MAVEBench, a benchmark of 50 film shots and 416 keyframes for evaluating camera motion, spatial layout, and character blocking. We propose MAVE, a video-to-blockout framework combining Metric observation, Agentic modeling, and evidence-based VErification. It first recovers cameras, human motion, and geometric evidence in a shared metric coordinate system. An agent then constructs an editable environment under constraints from these observations, while programmatic checks report inconsistencies to guide refinement. The outputs are editable Blender scenes and frame-aligned blockout videos. On MAVEBench, MAVE improves overall scores by 27.5-41.1 points across three agents over their corresponding vanilla baselines. With DeepSeek-1017, it achieves a score of 73.2 at an average API cost of only USD 0.054 per shot, outperforming the strongest vanilla baseline while reducing total reconstruction time by 53%. We also construct MAVEData-3K, a dataset of 3,000 blockout-video pairs for model training, and will release it publicly. Creators can change viewpoints and character paths, then render the edited scenes to condition video generation. Code is available at this https URL.
Comments: this https URL
Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Cite as: arXiv:2611.22464 [cs.CV]
  (or arXiv:2611.22464v1 [cs.CV] for this version)
  https://doi.org/10.48550/arXiv.2611.22464

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Yifan Chang [view email]
[v1] Tue, 17 Nov 2026 09:25:47 UTC (12,279 KB)

来源:arXiv cs.AI · arxiv.org