RecipeGen: A Step-Aligned Multimodal Benchmark for Real-World Recipe Generation

TL;DR

RecipeGen introduces a comprehensive, large-scale benchmark dataset for multimodal recipe generation tasks, addressing the lack of fine-grained alignment between recipe instructions, images, and videos in food computing.

Contribution

It provides the first real-world, multi-modal benchmark dataset with domain-specific evaluation metrics for recipe-based text-to-image, image-to-video, and text-to-video generation.

Findings

Benchmark results for T2I, I2V, T2V models

Domain-specific metrics for ingredient fidelity

Insights for future recipe generation models

Abstract

Creating recipe images is a key challenge in food computing, with applications in culinary education and multimodal recipe assistants. However, existing datasets lack fine-grained alignment between recipe goals, step-wise instructions, and visual content. We present RecipeGen, the first large-scale, real-world benchmark for recipe-based Text-to-Image (T2I), Image-to-Video (I2V), and Text-to-Video (T2V) generation. RecipeGen contains 26,453 recipes, 196,724 images, and 4,491 videos, covering diverse ingredients, cooking procedures, styles, and dish types. We further propose domain-specific evaluation metrics to assess ingredient fidelity and interaction modeling, benchmark representative T2I, I2V, and T2V models, and provide insights for future recipe generation models. Project page is available now.