Transfer Learning-Based Classification of Fresh and Stale Fruits and Vegetables Using Deep Convolutional Neural Networks

In Food supply chains and household consumption the spoilage of fruits or vegetables and their quality degradation is a major challenge. Early identification of fresh and stale produce can help reduce food waste and improve food safety. We use deep learning to automatically learn and classify the fresh and stale fruits and vegetables using image data. A publicly available fruit and vegetable dataset has been used for this research work. Three transfer learning models, namely MobileNetV2, ResNet50, and EfficientNetB0, were employed to perform multi-class image classification.