In this chapter, the performance of eXtreme Gradient Boosting Classifier (XGBClassifier) is compared with other classifiers for 2D object recognition. A fusion of several feature detector and descriptors (SIFT, SURF, ORB, and Shi Tomasi corner detector algorithm) is taken into consideration to achieve the better object recognition results. Various classifiers are experimented with these feature descriptors separately and various combinations of these feature descriptors. The authors have presented the experimental results of public datasets, namely Caltech-101 which is a very challenging image dataset. Various performance measures, i.e., accuracy, precision, recall, F1-score, false positive rate, area under curve, and root mean square error, are evaluated on this multiclass Caltech-101 dataset. A comparison among four modern well-known classifiers, namely Gaussian Naïve Bayes, decision tree, random forest, and XGBClassifier, is made in terms of performance evaluation measures. The chapter demonstrates that XGBClassifier outperforms rather than other classifiers as it achieves high accuracy (88.36%), precision (88.24%), recall (88.36%), F1-score (87.94%), and area under curve (94.07%) when experimented with the fusion of various feature detectors and descriptors (SIFT, SURF, ORB, and Shi Tomasi corner detector).