Comparative Analysis of Convolutional Neural Network (CNN) Architectures in Classification of Cattle and Pig Rambaks

Abstract

Rambak crackers are one of the food ingredients that have the characteristics of expansion and crispy texture. The general public often faces difficulties in distinguishing between pork and beef rambak crackers that have been processed, so it is important to rely on technology, especially artificial intelligence (AI), to help distinguish between them. This study was conducted to compare the capabilities of several CNN architectures in classifying images of pork and beef rambak crackers. The results of the study showed that the Xception architecture had the highest accuracy rate in classifying pork and beef rambak crackers, with an average accuracy rate of 98.24%.