The BRIEF descriptor has comparatively few bits, and can be computed using a set of intensity difference tests. Being a short binary descriptor, it has a low memory footprint, and the matching using this descriptor turns out to be very efficient with the Hamming distance metric. With BRIEF, the desired scale-invariance can be obtained by detecting features at different scales, although it does not provide rotation-invariance. The next code block demonstrates how to compute the BRIEF binary descriptors with scikit-image functions. The input images for matching used are the gray-scale Lena image and its affine transformed versions.
Let's now code this:
from skimage import transform ...
