FilmWorld: Agentic Novel-to-Film Generation through Dynamic Cinematic World Modeling

2026-07-21

AI Summary

This work addresses the challenge of automatically transforming novels into long-form, multi-scene films—a task hindered by the gap between literary abstraction and visual narrative coherence, which existing video generation models struggle to bridge due to difficulties in preserving entity states and causal consistency across scenes. The paper proposes the first two-stage dynamic world modeling paradigm tailored for novel-to-film generation: in the construction stage, textual content is concretized into persistent entities with explicit states and shot plans are formulated; in the evolution stage, visuals are generated via state anchoring, states are propagated across shots, and consistency is enforced through closed-loop validation. Built upon a multi-agent collaborative architecture, the system integrates narrative parsing, state-driven shot planning, and cross-scene state propagation, and introduces FilmEval—the first evaluation framework for long-video generation. Evaluated on a benchmark of 15 novels stratified by difficulty, the method significantly outperforms existing approaches in narrative fidelity and cross-scene consistency.

Abstract

Translating novels into films poses a grand challenge for generative artificial intelligence, requiring conversion of abstract literary prose into long-form, multi-scene visual narratives. While current video generation models excel at short, single-scene clips within narrow temporal and spatial contexts, novel-to-film generation operates in a more complex regime, demanding long-duration content across diverse scenes with dynamically evolving entity states. To address this, we formalize novel-to-film generation as dynamic cinematic world modeling, decomposed into two phases: construction, which grounds abstract, underspecified literary narratives into concrete, stateful, and persistent world entities; and evolution, which governs how these entities dynamically update under plot progression to maintain causal consistency across scenes. We propose FilmWorld, an end-to-end agentic system where two groups of specialized agents collaborate to instantiate these phases. Construction-side agents perform narrative structured translation, world entity state modeling with visual anchoring, and state-driven shot planning, progressively projecting literary language into a cinematic blueprint. Evolution-side agents perform state-anchored visual generation, cross-shot dynamic state propagation, and closed-loop state verification to maintain causal consistency and visual coherence. To address the evaluation gap in long-form generation, we introduce FilmEval, a systematic evaluation framework that couples a difficulty-graded benchmark of 15 representative novels with an automated protocol of nine objective metrics spanning three dimensions: cinematic presentation, film consistency, and novel fidelity. Experiments demonstrate that FilmWorld consistently outperforms state-of-the-art video generation agent systems, with particularly pronounced improvements in narrative fidelity and cross-scene consistency.

Problem

• novel-to-film generation
• dynamic cinematic world modeling
• long-form video generation
• cross-scene consistency
• narrative fidelity

Innovation

• dynamic cinematic world modeling
• agentic video generation
• stateful entity propagation
• novel-to-film translation
• FilmEval