Machine Learning for OpenCV

Using the Scale-Invariant Feature Transform (SIFT)

However, corner detection is not sufficient when the scale of an image changes. To this end, David Lowe came up with a method to describe interesting points in an image, independent of orientation and size. Hence, the name scale-invariant feature transform (SIFT).

In OpenCV 3, this function is part of the xfeatures2d module:

In [6]: sift = cv2.xfeatures2d.SIFT_create()

The algorithm typically works in two steps:

• Detect: This step identifies interesting points in an image (also known as keypoints)
• Compute: This step computes the actual feature values for every keypoint

Keypoints can be detected with a single line of code:

In [7]: kp = sift.detect(img_bgr)

In addition, the drawKeypoints ...