In the previous post on image feature matching using OpenCV, we looked at how distinct features of an image form keypoints and descriptors, and how similarities between images can be matched using these keypoints. We looked at two detecting methods: ORB and SIFT.
But the question is, which one is better of them in practice? Spoiler alert: Like everything else, there is no clear winner and it depends on your application and you need to test to find out which one works well for your use-case. So let us test and evaluate both the detectors. In this blog, we are also going to augment the images and then run the detectors to check their robustness.
