Advanced deep learning algorithms in food quality and authenticity

Abstract

Ensuring food quality and authenticity is critical to the food industry and consumers. With the advancements in Industry 4.0 and Artificial Intelligence (AI) technologies, deep learning (DL) offers unparalleled opportunities to extract information and make decisions on complex or large datasets. However, conventional convolutional neural networks (CNN) and recurrent neural networks (RNN) have limitations. The development of advanced DL algorithms can accommodate the growing demand for complex tasks and herald revolutionary breakthroughs in the field of food quality and authenticity identification, which will continue to be driven by the ongoing development of advanced DL. This review provides a comprehensive overview of various advanced DL algorithms for food quality identification and food authenticity analysis, including advanced variants of CNN, lightweight DL, sequential neural networks, graph neural networks (GNN), deep generative models (DGM), and target detection algorithms. It also surveys recent applications of advanced DL algorithms for Food quality inspection and authenticity analysis. This review discusses the challenges associated with advanced DL and the future trends, offering new insights into the development of advanced DL algorithms in food quality and authenticity. Challenges such as overfitting, scalability, interpretability, accessibility, data privacy, algorithmic bias, and the creation of large databases must be addressed in the application of advanced DL algorithms to drive their further iterations.