Answers are grounded in this paper and cite their sources.
Global food waste has increased due to inefficiencies in production, processing, and distribution, making food security an increasingly critical issue. Rising global food demand and substantial losses along the food supply chain emphasize the need to optimize available resources for effective food management. Spoilage detection matters because a rotten fruit left among sound ones can spread decay and cause further losses. Sorting is still largely done by hand, which is slow, costly, and inconsistent between inspectors. This study proposes a deep learning method that identifies the fruit type and classifies it as fresh or rotten from a single image, aiming to reduce manual sorting effort and cost. Convolutional Neural Networks (CNNs), with their ability to automatically extract hierarchical features from images, are well-suited for distinguishing between fresh and rotten fruits. In our approach, convolutional layers detect patterns and edges, while pooling layers reduce feature dimensionality to improve classification efficiency. The enhanced CNN, trained with data augmentation on 13,568 images of fresh and rotten apples, bananas, and oranges (six classes), identifies the fruit type and predicts its freshness in one pass. Experimental results show a validation accuracy of 99.03% and a training accuracy of 99.46%, demonstrating the effectiveness of the proposed method. The results highlight the potential of deep learning-based automated systems in agricultural and retail settings, where they can improve fruit quality assessment, ensure consumer safety, and reduce operational costs. This work underscores the importance of integrating machine learning technologies into food production and supply chains to enhance efficiency and sustainability.
No takes yet. Share an insight, caveat, or question.
Idrees et al. (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: