AI řízená detekce alergenů v potravinách a doporučení stravy využívající velké jazykové modely a počítačové vidění

Abstract

This bachelor thesis deals with the design and implementation of a mobile application for detecting food allergens from image and text inputs. The application enables the analysis of photographs of meals, menus, product labels, as well as text-based food names. The solution is implemented using the .NET MAUI framework, which supports the development of cross-platform mobile applications. A key component of the application is the integration of a multimodal large language model provided by OpenAI through an API interface. The large language model is used to identify potentially hazardous ingredients and to interpret the presence of allergens based on the provided input data. Given the wide potential user base, emphasis is placed on the intuitiveness and simplicity of the user interface. The result is a functional prototype of the application, which was experimentally verified using real-world data and also underwent user testing. In this way, not only the correct functionality but also the practical benefits of the proposed solution were confirmed.