Local Feature Detectors Performance Analysis on Digital Image

The article puts an experiment on application of widely used ORB, SIFT and SURF feature point detectors represented by the corresponding functions in the OpenCV library, to the images of the most common object classes such as human faces, fine details and artificial images. The considered detectors display as a result a huge number of points that are not classified or structured. Building an appropriate classifier would greatly increase the efficiency of subsequent image processing operations: localization, recognition, search, and tracking of objects. The article analyzes the effectiveness of the experimental results at a quantitative and qualitative level taking into account the conditions and limitations (primarily temporal) on solving practical problems in the era of Big Data, as well as taking into account the fact that some detectors are proprietary. According to the analysis results of the usage effectiveness of the features points detectors considered in the work the practical recommendations for specific use cases are given at the end of the work.