Classification of Chicken Meat Freshness Based on YCbCr Color and Fractal Features Using KNN Method

Abstract

The assessment of chicken breast meat freshness is crucial for ensuring food safety and meeting consumer expectations in the poultry industry. This study focuses on classifying chicken breast meat freshness into ‘fresh’ and ‘not fresh’ categories using a dataset of 349 images. Initially, images were segmented based on the YCbCr color space, followed by fractal dimension feature extraction using the box counting method. Utilizing a K-Nearest Neighbors (KNN) classifier, the dataset was validated using the 10-fold cross-validation method, achieving a classification accuracy of 94.55%, with precision and recall at 96% and 93.07%, respectively. The integration of YCbCr color segmentation and fractal dimension feature extraction proved effective in distinguishing freshness levels of chicken meat, offering a reliable and objective approach for assessing chicken meat freshness with implications for poultry industry quality control. This research contributes to advancing image-based freshness classification techniques, highlighting the potential of combining color analysis and fractal features for accurate and efficient results.

Keywords; Fractal Dimension, KNN, Meat freshness, YCbCr