Classification of Chicken Meat Freshness Using Support Vector Machine and Hue Saturation Intensity

Abstract

Chicken meat is a popular source of animal protein in Indonesia due to its high nutritional value, affordable price, and easy processing. The identification of chicken meat freshness is currently still done manually through visual or tactile inspection, but this method has limitations, especially if consumers are less skilled in distinguishing the quality of chicken meat freshness. Therefore, an automated system is needed to classify the freshness level of chicken meat based on images. This research aims to develop an image processing system in classifying the freshness level of chicken meat by utilizing the Support Vector Machine (SVM) method with Hue Saturation Intensity (HSI) based color feature extraction. This process is done by converting the RGB image into HSI, then extracting the Hue, Saturation, and Intensity values and classifying using a polynomial kernel. This study used 450 chicken meat images, with 360 training data and 90 test data. The developed system successfully achieved an accuracy of 65.56%. The test results show that the system is reliable in classifying the freshness level of chicken meat. This system has the potential to support the identification of meat freshness efficiently and objectively, while at the same time improving food safety.