Design and Development of Vegetarian Food Identifier Application

Abstract

Vegetarianism, often practiced for religious or health reasons, is frequently overlooked by businesses, making it difficult for vegetarians to verify and obtain pre-packaged foods and access shared knowledge. This study aims to develop an Android application that classifies pre-packaged food products as vegetarian or non-vegetarian using Optical Character Recognition (OCR) and deep learning, while also supporting community engagement through knowledge sharing. The system follows the Waterfall model with Royce’s iterative feedback, utilizing React Native, Laravel, MySQL, and FastAPI in a microservice architecture; OCR extracts text from ingredients list, which are processed by a deep learning AI model. Results show the application successfully simplifies food verification and supports user interaction through posting, commenting, and liking features. While the application achives its objectives, limitations include Android-only support, UI simplicity, and OCR sensitivity; future work may include iOS or web versions, dynamic label management, and improved AI model accuracy through domain-specific fine-tuning.