Comparison of popular features used in object recognition

In this research, a comparison of popular features used in object recognition was made. The development of technology has brought innovations in computer vision tasks. Computer vision and object detection are currently used in autonomous vehicles, face recognition, medical imaging and many other areas. Three popular features and two data sets used in object recognition were used in the study. The performances of the features in the data sets and the effects of different data sets on the feature quality were examined. According to the results of the examination, the feature qualities varied in different data sets. The data sets of this study were determined as VisDrone data set and MNIST data set. The feature extraction methods were determined as HOG, SIFT and ORB feature extractors. SIFT feature extractor exhibited the best performance in VisDrone data set. ORB feature extractor had the best performance in MNIST data set.