BRISK – scale invariant image matching

BRISK is another detector available in Julia. The most significant benefit of BRISK is that it is scale and rotation invariant. Scale invariance comes with computational costs, which makes it slightly slower than ORB.

BRISK also allows using any other keypoints descriptor provided by, for example, imcorner, such as harris. We continue using FAST as in the preceding examples.

So, let's start using BRISK. As always, we start by loading the packages and images, as shown in the following code:

using Images, ImageFeatures, CoordinateTransformations, ImageDraw, ImageView

img1 = Gray.(load("sample-images/cat-3417184_640.jpg"))
img2 = Gray.(load("sample-images/cat-3417184_640_watermarked.jpg"))

We have set the target to demonstrate that BRISK is scale and rotation invariant. Then, we apply two types of transformations...