Research on meat adulteration based on one-dimensional convolutional neural network (1DCNN) combined with electrochemical technology

This study presents a novel methodology for detecting meat adulteration by integrating electrochemical sensing with a one-dimensional convolutional neural network (1DCNN). Addressing the challenge of accurately identifying adulterants in various meats, including beef, lamb, pork, chicken, and particularly lard due to religious dietary constraints, the research focuses on both raw and cooked meat samples. The developed approach uses disposable screen-printed carbon electrodes for electrochemical measurements, capturing unique “electrochemical fingerprints” of the meats. The 1DCNN models, optimized for binary classification problems, demonstrate exceptional accuracy, achieving test accuracies ranging from 95.2 to 98.1% across different meat types. Sensitivity and specificity of the models consistently exceed 95%, signifying a robust detection capability. The study’s findings indicate that the electrochemical-1DCNN fusion can detect adulteration levels as low as 10%, showcasing its potential as a rapid and reliable method for meat authentication in the food industry.