Classification of Gamebird Images Using Color Histogram Intersection

Abstract

Gamebirds and cockfighting is a multi-billion industry in the Philippines and is a viable research area. This study explores the classification of gamebird images using color histogram intersection. The dataset of gamebird images were taken from Pangasinan Breeders’ Cup 2018 derby fights and are classified into dom, grey, red, and white classes. Color histograms were extracted from these images and experiments on the number of models and channels to be utilized were conducted, then the performance of the method was assessed. The results show that the classification of gamebird images using color histogram intersection performed best in the single model using all channels set-up with sensitivity of 89.75%, specificity of 96.58%, and accuracy of 94.88%.