From Bytes to Bites: Revolutionizing Culinary Creativity Through Deep Learning and AI-Generated Recipes

In the field of computer vision, pre-trained object detection models are highly adept at recognizing specific classes of objects within an image. This chapter presents a novel application that harnesses the power of a state-of-the-art object detection model, YOLOv8, to create a new model that can accurately detect and identify a wide variety of 26 vegetables. During the development of the model, we employed the well-established squared bounding box mechanism to precisely locate each vegetable within the image. Additionally, we significantly improved the application's functionality by integrating a large language model (LLM) at the final stage. This integration enables users to generate personalized culinary instructions based on the detected vegetables in their preferred language, complete with audio narration and the ability to select the desired number of recipes. The newly developed model has demonstrated remarkable accuracy, attesting to its efficacy in vegetable identification tasks.