Adulteration detection in minced beef using low-cost color imaging system coupled with deep neural network

Major processed meat products, including minced beef, are one of the favorite ingredients most people because they high in protein, vitamins, and minerals. The demand prices make products vulnerable to adulteration. In addition, eliminating morphological attributes makes authenticity beef challenging identify with naked eye. This paper aims describe feasibility study adulteration detection using a low-cost imaging system coupled deep neural network. proposed method was expected be able detect There were 500 captured images samples. Then, there 24 color textural features retrieved from image. samples then labeled evaluated. A network (DNN) developed investigated support classification. DNN also compared six machine learning algorithms form accuracy, precision, sensitivity feature importance analysis performed obtain impacted classification results. model accuracy 98.00% without selection 99.33% selection. has best performance individual up 99.33%, precision 98.68%, 98.67%. work shows enormous potential application rapidly adulterants performance.