Student : MJ Maotwana
About the student
I’m an Honours Computer Science student who loves working with data and solving problems. I enjoy using technology in creative ways to make life easier and find simple, practical solutions to everyday challenges.
About the Project
My project is about creating an AI system that can tell how fresh fruits and vegetables are just by looking at pictures. Food waste is a big problem in South Africa and around the world. Many people throw away food that is still good because they are not sure if it is fresh, which leads to waste, money loss, and food shortages. The system uses computer vision and deep learning to classify produce as fresh or spoiled. Different machine learning models were tested, including Logistic Regression, Support Vector Machines (SVM), and a pretrained Convolutional Neural Networks (CNN) model called ResNet18. The ResNet18 model performed the best, reaching 97% accuracy, compared to 92% for SVM and 74% for Logistic Regression. The system is designed as a web app, making it easy for households, stores, and supply chain users to check produce quality. Shoppers can take a photo of fruits or vegetables to see if they are fresh, while stores can use it to manage stock and reduce waste.
