Search papers, labs, and topics across Lattice.
This paper introduces ReDesign, an innovative framework that recovers editable design structures from raster images by employing an agentic approach to build a hierarchical layer system through the selective composition of specialized tools. The framework incorporates a verification mechanism that mitigates error accumulation during the decision-making process, ensuring reliable outputs despite the challenges of imperfect tool performance. Evaluation against the Figma Edit Replay Benchmark demonstrates that ReDesign achieves superior visual fidelity and editability, surpassing existing methods in handling layout, color, and text modifications.
ReDesign not only recovers editable design files from images but does so with unprecedented accuracy and flexibility, setting a new standard for design workflows.
Recovering an editable design file from a raster image is a common and costly bottleneck in modern design workflows, yet remains challenging since editability depends on recovering multi-modal attributes, such as typography, vector geometry, colors, grouping, and layer ordering. We present ReDesign, an agentic framework that grows an editable layer hierarchy by selecting and composing specialized tools across modalities. To keep this long decision process reliable despite imperfect tool outputs, we introduce graceful verification at each expansion, which provides local accept, prune, or retry feedback that prevents error accumulation and avoids large scale reruns. To evaluate editability at scale, we introduce the Figma Edit Replay Benchmark, consisting of 909 raw Figma files and 14,796 controlled edit instructions that replay edits on reconstructed outputs. Across this benchmark and standard reconstruction metrics, ReDesign achieves strong visual fidelity while delivering the highest editability across layout, color, and text edits, outperforming layered decomposition baselines and serial tool use pipelines.
