Search papers, labs, and topics across Lattice.
MangaFlow is an agentic framework that decomposes manga generation into distinct steps: planning, grounding, layout, rendering, composition, and text placement, enabling controllable long-form manga creation. It introduces explicit intermediate variables for layout and visual references, allowing for user-controlled refinement of panel geometry, references, and text. Experiments demonstrate improved layout adherence and cross-panel consistency compared to direct generation baselines, while also supporting flexible human control.
Finally, a framework that disentangles the messy process of manga creation, giving you knobs to tweak layout, characters, and text across entire stories.
End-to-end manga generation is a structured visual storytelling task that requires story decomposition, recurring character and scene grounding, page layout design, panel rendering, page composition, and lettering. However, existing generative models often perform direct page synthesis, entangling these factors in a single visual output and limiting precise control over layout geometry, visual references, and cross-panel consistency. To address these limitations, we propose MangaFlow, an agentic framework for controllable long-form manga generation that decomposes manga creation into planning, grounding, layout construction, reference-conditioned rendering, composition, and text placement. By treating layout and visual references as explicit intermediate variables, MangaFlow enables both simple text-to-manga generation and more precise user-controlled manga creation. This design exposes layout, visual assets, and lettering as editable intermediate controls for refining panel geometry, references, and text placement. To support long-form consistency, MangaFlow introduces a story section memory that links section descriptions with corresponding character, scene, and object references for reuse across panels. We further present a meta-benchmark for evaluating layout controllability, visual consistency, and generation quality. Experiments show that MangaFlow improves layout adherence and cross-panel consistency over direct generation baselines while supporting flexible human control.
