---
title: "PACR-Video：参数高效的多镜头长视频外推框架"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2607.06481v1"
permalink: "https://feed.astralor.com/items/itm_8468b1d27b83080f4237823b"
published: "2026-07-07T16:41:25.000Z"
score: 78
channel: "tech"
tags: ["PACR-Video","低秩时序适配器","参数高效微调","扩散变换器","提示路由","长视频生成"]
---

# PACR-Video：参数高效的多镜头长视频外推框架

## AI 点评

> 这篇论文值得关注，因为它解决了长视频生成中一直悬而未决的‘微调成本 vs 连贯性’矛盾。对于从事视频生成、扩散模型或参数高效微调的研究者来说，PACR-Video 的思路极具启发——将叙事结构化为可路由的提示模块，而不是无差别地记忆或微调全部参数。你可能不需要巨大算力就能在自己的模型上尝试类似设计。

## 速览

PACR-Video 是一种参数高效的多镜头长视频外推框架，它冻结文本到视频扩散变换器，仅通过低秩时序适配器和提示路由实现连贯的长视频生成。该方法采用递归提示库存储实体、动作和风格信息，并通过适配器门控根据叙事依赖路由。在六个多镜头和长视频基准上，PACR-Video 在分布质量、语义对齐、身份一致性、时间平滑性、运动稳定性、过渡连贯性和人类偏好上均超越现有基线，展示了紧凑提示路由和轻量时序适配的足够可控能力。

## 关键点

- 冻结T2V扩散Transformer，仅训练低秩时序适配器和提示路由，实现参数高效。
- 递归提示库存储多镜头间的实体、位置、动作和风格信息，通过门控路由维持长程连贯性。
- Shor-Local/Story-Global 联合训练目标结合下一镜头重建、跨镜头身份对比和提示稀疏正则化。
- 在六个基准上全面超越现有方法，包括流式、递归和记忆增强基线。

## 全文

We present PACR-Video, a parameter-efficient framework for multi-shot long video extrapolation that preserves recurring entities, scene structure, visual style, and causal progression without full generator fine-tuning. PACR-Video keeps a text-to-video diffusion transformer frozen and augments it with low-rank temporal adapters conditioned by learned shot-role prompt tokens. To maintain long-horizon coherence, it builds a recursive prompt bank that stores compact entity, location, action, and style prompts from previous shots, then routes them through adapter gates according to predicted narrative dependencies. A Shot-Local/Story-Global tuning objective combines next-shot reconstruction, cross-shot identity contrast, and prompt sparsity regularization, while an adapter composition schedule balances early-shot visual consistency with later-shot event progression and viewpoint change. Across six multi-shot and long-video benchmarks, PACR-Video outperforms text-to-video, tuning-based, memory-augmented, streaming, and recursive-context baselines on distributional quality, semantic alignment, identity consistency, temporal smoothness, motion stability, transition coherence, and human preference. These results show that compact prompt routing and lightweight temporal adaptation provide sufficient controllable capacity for stable long video extrapolation.

---

引用前请核对原文;本文由星际流动整理:https://feed.astralor.com/items/itm_8468b1d27b83080f4237823b