ShowTable: Unlocking Creative Table Visualization with Collaborative Reflection and Refinement
Zhihang Liu, Xiaoyi Bao, Pandeng Li, Junjie Zhou, Zhaohe Liao, Yefei He, Kaixun Jiang, Chen-Wei Xie, Yun Zheng, Hongtao Xie
2025-12-17
Summary
This paper introduces a new challenge for AI image generation: turning data tables into visually appealing and accurate infographics. It then presents a system called ShowTable that uses a combination of AI techniques to accomplish this.
What's the problem?
Current AI models are really good at making general images, but they struggle when you need them to *understand* data and represent it visually in a specific, correct, and creative way. Think about taking a spreadsheet and turning it into a clear and beautiful chart – that requires more than just general image-making skills; it needs reasoning and precise mapping of data to visuals, something existing models lack.
What's the solution?
The researchers developed ShowTable, which works in stages. First, a powerful AI called an MLLM (Multi-Modal Large Language Model) analyzes the table and figures out a plan for the infographic. Then, another AI, a diffusion model, actually creates the image based on the MLLM’s instructions. The MLLM doesn’t just give instructions once; it checks the image, finds errors, and then gives the diffusion model *refined* instructions to fix them. This back-and-forth process continues until the infographic is accurate and looks good. They also created new ways to automatically generate training data for these AIs and a benchmark to test how well they perform.
Why it matters?
This work is important because it pushes AI beyond simply creating pretty pictures. It tackles a real-world problem – data visualization – that requires genuine understanding and reasoning. If AI can reliably turn data into clear visuals, it could help people understand complex information more easily in fields like science, business, and education.
Abstract
While existing generation and unified models excel at general image generation, they struggle with tasks requiring deep reasoning, planning, and precise data-to-visual mapping abilities beyond general scenarios. To push beyond the existing limitations, we introduce a new and challenging task: creative table visualization, requiring the model to generate an infographic that faithfully and aesthetically visualizes the data from a given table. To address this challenge, we propose ShowTable, a pipeline that synergizes MLLMs with diffusion models via a progressive self-correcting process. The MLLM acts as the central orchestrator for reasoning the visual plan and judging visual errors to provide refined instructions, the diffusion execute the commands from MLLM, achieving high-fidelity results. To support this task and our pipeline, we introduce three automated data construction pipelines for training different modules. Furthermore, we introduce TableVisBench, a new benchmark with 800 challenging instances across 5 evaluation dimensions, to assess performance on this task. Experiments demonstrate that our pipeline, instantiated with different models, significantly outperforms baselines, highlighting its effective multi-modal reasoning, generation, and error correction capabilities.