Julia Yili Shang¹ and Julian Avellaneda², ¹USA, ²California State Polytechnic University, USA

Previous research demonstrated that an average American adult could easily over consume the daily suggested amount of added sugar. Our team realized that a typical dessert contains way more sugar than suggested amount, which can be potentially harmful to obese and diabetic patients. We decided to develop a mobile application scans conventional baking recipe, recognize the and generates a healthier recipe reduced in sugar using AI to replace some of the ingredients to healthier ones. The three major technical components are authentication service, AI, and text recognition and matching function. We integrated OpenAI to build the template, Spacy function and image.scan for text recognition, and firebase authentication for login functions. Two experiments were conducted to test the accuracy of AI and text recognition. All the testing's suggested that the app is able to adjust any baking recipe to a healthier alternative, while guaranteeing the basic tastes.

Diabetes, Baking, Mobile Application, Health