Machine learning combined with electrochemical sensor for rapid detection of Sudan Red I in food

Xi Sun,Feng Liu,Xiao Xue

DOI: https://doi.org/10.1007/s11694-023-02150-w

2023-09-28

Abstract:Ensuring food safety necessitates rapid and reliable food contaminant detection methods. Here, we present a novel approach that integrates electrochemical sensors with machine learning to detect Sudan Red I in food. Our study showcases a modified stainless steel needle (SSN) electrode adorned with silver nanoparticles (AgNPs) and investigates three distinct convolutional neural network (CNN) architectures: Inception V1, ResNet-50, and SqueezeNet V1.1. This combination aims to predict the sensor response accurately. The AgNPs/SSN electrode displays heightened electrochemical activity towards Sudan Red I, offering a linear detection range of 0.1–20 μM and an exceptional 10 nM detection limit. Notably, the Inception V1 and ResNet-50 networks exhibit superior predictive capability compared to SqueezeNet V1.1. Our interdisciplinary approach establishes a rapid, sensitive, and dependable detection system for food contaminants. By fusing electrochemical sensors with machine learning, we propel the development of intelligent detection systems with significant implications for food safety enhancement.Graphical abstract

food science & technology