Computer Vision with OpenCV 3 and Qt5

Extracting and matching descriptors

A descriptor in computer vision is a way of describing a keypoint that is completely dependent on the specific algorithm that is used for extracting it, and, unlike keypoints (which are defined in the KeyPoint class), descriptors do not have a common structure, except more or less the fact that each descriptor represents a keypoint. Descriptors in OpenCV are stored in Mat classes, where each row in the resulting descriptor Mat class refers to a descriptor of a keypoint. As we learned in the previous section, we can use the detect function of any FeatureDetector subclass to basically detect a set of keypoints from an image. Similarly, we can use the compute function of any DescriptorExtractor subclass to ...