Object Classification Techniques using Tree Based Classifiers

Abstract

Object recognition is presently one of the most active research areas in computer vision, pattern recognition, artificial intelligence and human activity analysis. The area of object detection and classification, attention habitually focuses on changes in the location of anobject with respect to time, since appearance information can sensibly describe the object category. In this paper, feature set obtained from the Gray Level Co-Occurrence Matrices (GLCM), representing a different stage of statistical variations of object category. The experiments are carried out using Caltech 101 dataset, considering sevenobjects viz (airplanes, camera, chair, elephant, laptop, motorbike and bonsai tree) and the extracted GLCM feature set are modeled by tree based classifier like Naive Bayes Tree and Random Forest. In the experimental results, Random Forest classifier exhibits the accuracy and effectiveness of the proposed method with an overall accuracy rate of 89.62%, which outperforms the Naive Bayes classifier.