Ensemble Deep Learning for Intelligent Ingredient Detection and Personalized Recipe Recommendation

Abstract

Making meals and dishes that promote a healthy lifestyle is a challenge for many people these days. This project makes cooking easier with an AI-powered system that uses photos to identify ingredients and suggests customized recipes. Users upload a picture of the ingredients they have on hand, and the system uses an ensemble model that combines Faster R-CNN and YOLOv8 to accurately identify them. The system guarantees accurate ingredient detection by combining the detections from both models using NMS Fusion and the IoU metric.Additionally, it takes into account preparation time, dietary restrictions, and allergies, customizing recipe suggestions to suit personal preferences.  This automatic method, which learns user preferences over time to provide even better choices, saves time and effort compared to traditional recipe searches.Combining ingredients wisely increases meal possibilities, minimizes food waste, and encourages better eating practices.Future developments might include voice assistants, linguistic assistance, and smart kitchen connectivity, which would make the system even more user-friendly and accessible for home cooks everywhere.