Seeing is Improving: Visual Feedback for Iterative Text Layout Refinement
This work improves layout generation for design-oriented multimodal large language models, offering a solution to a specific bottleneck in automated visual design.
The paper tackles the problem of automated layout generation from natural language descriptions by addressing the lack of visual feedback in existing code-only methods, which often compromises readability and aesthetics. The proposed Visual Feedback Layout Model (VFLM) uses reinforcement learning with a visually grounded reward model to iteratively refine layouts, outperforming advanced MLLMs, existing layout models, and code-only baselines across multiple benchmarks.
Recent advances in Multimodal Large Language Models (MLLMs) have enabled automated generation of structured layouts from natural language descriptions. Existing methods typically follow a code-only paradigm that generates code to represent layouts, which are then rendered by graphic engines to produce final images. However, they are blind to the rendered visual outcome, making it difficult to guarantee readability and aesthetics. In this paper, we identify visual feedback as a critical factor in layout generation and propose Visual Feedback Layout Model (VFLM), a self-improving framework that leverages visual feedback iterative refinement. VFLM is capable of performing adaptive reflective generation, which leverages visual information to reflect on previous issues and iteratively generates outputs until satisfactory quality is achieved. It is achieved through reinforcement learning with a visually grounded reward model that incorporates OCR accuracy. By rewarding only the final generated outcome, we can effectively stimulate the model's iterative and reflective generative capabilities. Experiments across multiple benchmarks show that VFLM consistently outperforms advanced MLLMs, existing layout models, and code-only baselines, establishing visual feedback as critical for design-oriented MLLMs. Our code and data are available at https://github.com/FolSpark/VFLM.