AUTHOR(S): 

Ibrahim El-Henawy, Hazem. M. El bakry, Hagar M. El Hadad, Nikos Mastorakis

TITLE

Muzzle Feature Extraction Based on gray level co-occurrence matrix

KEYWORDS

bovine classification; average filter; Median filter; gray level co-occurrence matrix (GLCM)algorithm; Naive Bayes and Decision tree.

ABSTRACT

Nowadays veterinarians pay a numerous effort to save bovines' products because of its rapid growth. The critical point in this paper is to classify and distinguish between large different groups of bovines with high degree of accuracy. This paper presents two bovine's classification models depending on J48 decision tree and Naive Bayes classifier. These two models consist of three phases; pre-processing phase, texture feature extraction phase and classification phase. Pre-processing phase consists of average filter and median filter to remove noise from muzzle image. The second phase used gray level co-occurrence matrix (GLCM) algorithm to extract image features. Then the last phase used decision tree in the first model and used Naive Bayes the second model to classification muzzles and comparing the accuracy result. The used data base consists of fifty-two different bovine. The experimental result proves the advancement of decision tree classifier than Naive Bayes classifier. By comparing the result of decision tree with Naive Bayes the accuracy rate in case of using fifty-two different groups is 89.64% comparing with 75.38% in case of using Naive Bayes classification system..

Cite this paper

Ibrahim El-Henawy, Hazem. M. El bakry, Hagar M. El Hadad, Nikos Mastorakis. (2016) Muzzle Feature Extraction Based on gray level co-occurrence matrix. International Journal of Veterinary Medicine, 1, 16-24