Research Highlights

This research project develops innovative technologies to enhance accessibility, usability, and cross-cultural exchange of cooking recipes through artificial intelligence. It presents three complementary systems designed to transform how recipes are translated, visualized, and generated. TransRecipe enables automatic translation of Chinese recipes into English without human intervention, allowing users to explore diverse culinary traditions beyond language barriers. VisualRecipe converts text-based recipes into visual formats, including images and videos, thereby improving comprehension and usability for individuals who prefer visual learning or encounter challenges with written instructions. GenRecipe further extends this capability by applying deep learning techniques to generate complete recipes from cooking videos, accurately identifying ingredients, preparation steps, and cooking processes.

Building on these advancements, the project establishes an integrated framework that significantly improves recipe accessibility for a broad audience, including both home cooks and professional chefs with varying levels of expertise. Through the integration of machine translation, computer vision, and artificial intelligence, it introduces a novel approach to culinary knowledge sharing. The outcomes enhance user engagement in recipe discovery and preparation, while also supporting innovation in digital gastronomy and intelligent content creation.