Image matching is a frequently performed task in computer vision with numerous applications. Keypoint features such as SIFT, SURF, and ORB have been widely used in image matching and provides robustness against up to a significant degree of scaling, rotation, and blurring. The proportion of keypoints matched between two images can be used as a metric for how much the two images are related to each other. The image matching problem can be viewed as a classification problem in which the output is one of two classes, matching or non-matching. Ensemble methods, such as boosting and bagging, combine multiple child classifiers to achieve better accuracy than any of those classifiers alone. In this study, a method combining the keypoint features SIFT, SURF, and ORB with Adaptive Boosting is proposed to try to achieve a higher accuracy. Bag of Visual Words is applied to limit the computational cost of matching SIFT and SURF features. Results show that the ensemble model outperforms each keypoint feature alone. In the future, more image keypoint features can be added and different ensemble methods experimented with to try to further improve the model's performance.