International Journal of Science, Engineering and TechnologyAuthors: Varalakshmi S P, Banibrata Paul
Abstract: Modern kitchens face challenges in meal preparation due to lack of ingredient planning, leading to repetitive meals or food wastage. Bitezy is a real-time, AI-powered food recommendation system that takes user input in the form of images, text, or both, and suggests personalized recipes using available ingredients. The system uses a deep learning based object detection model, YOLOv5, to identify ingredients from uploaded images. To match these ingredients with recipes, Bitezy leverages a Word2Vec-based semantic similarity model that retrieves contextually relevant matches. The Roboflow dataset containing over 43 labelled ingredient categories is used for training the detection model. Bitezy also incorporates filters such as cuisine type, meal course, and diet preferences to further customize outputs. The output is comprehensive—providing the recipe title, image, ingredient list, instructions, estimated cooking time, servings, and nutritional values. Experimental evaluation showed that the system achieves high accuracy, fast response time, and robust performance in both single and multi-modal inputs. Bitezy offers a scalable, accessible, and intelligent platform for modern meal planning, contributing to sustainability and better food management. Its intuitive interface ensures ease of use for users of all ages. The system's modular architecture enables easy future integration with voice-controlled assistants and smart kitchen devices. The system promotes dietary awareness and healthy eating habits through personalized suggestions. Bitezy bridges the gap between technology and everyday cooking, Revolutionizing the way users engage with their kitchen environment.
DOI:
