Abstract

This project proposes an AI-powered Android application for evaluating fruit quality, with the potential to extend to assessing the health of vegetables as well. This tool offers buyers and sellers a valuable resource for examining the quality of fruits and vegetables. Leveraging one of the most efficient and lightweight deep learning models, EfficientNetB5, the application performs real-time quality assessments on fruits such as apples, bananas, and oranges. Each item is classified with a rating of "Good," "Bad," or, in some cases, a "Mixed" label, indicating that the model detected both positive and negative visual cues from the image.
The application, designed for Android devices, is accessible and user-friendly. Users can upload an existing image or capture a new one within the app, which then analyzes the image to evaluate freshness and quality. For consumers, this facilitates better purchasing decisions, reducing waste and increasing satisfaction. For retailers and the food industry, it provides an affordable, automated solution, with the potential to incorporate a robotic arm to enhance product quality through automation.
This project serves as a progressive example of AI integration into daily life, demonstrating how artificial intelligence can transform consumer behavior and reduce environmental impact through improved decision-making and automation.