Efficient Fruits Freshness Classification using Quantum Convolutional Neural Networks (Q-DCNN)
Maddigatla Hemalatha, Suneel Kumar Budithi, Venkata Krishna Rao Likki +3 more
In agriculture and food processing, sorting fruits by how fresh they are is an important job. It helps eliminate the need for manual inspections and stops losses after harvest. Quantum Deep Convolutional Neural Networks (Q-DCNN) have become a potential way to do high-accuracy visual classification as quantum-inspired machine learning has changed quickly. This paper presents an effective Q-DCNN-based system for autonomous fruit freshness classification utilizing a fully end-to-end CNN architecture devoid of transfer learning. The methodology commences with structured data preprocessing, encompassing color normalization, image scaling, augmentation, and labeling to guarantee robustness across various fruit classifications. The suggested Q-DCNN uses quantum-inspired techniques like superposition-based feature encoding and entanglement-oriented convolutional mixing to improve feature representation while keeping the computing footprint small. The network has a lot of convolutional, pooling, and fully connected layers. It learns from scratch how to find freshness features that are relevant to a task. In the last step, a SoftMax classifier is employed to guess the class. The suggested Q-DCNN has been tested dataset and the results show that it gets 99% accuracy, respectively. The model is also quite efficient, taking only 6 ms on average to produce the final classification result. These results show that Q-DCNN is a reliable, low-latency, and very accurate way to check the freshness of fruit in real time.
International conference Computer Networks and Inventive Communication Technologies•May 11, 2026•0 citations