Investigate performance optimization in `_corner_fast` about scikit-image

[spdfghi]

@lagru i made a new pr #7394
I just made a test of time cost by corner_fast:
this pr: 1.35s and brefore: 3.22s

here is my test code as you provided before, and simply used a large image input.
import numpy as np
import timeit
from skimage.feature import corner_fast, corner_peaks
square = np.zeros((1200, 1200))
square.astype(int)
def test():
corner_fast(square)

execution_time = timeit.timeit(test, number=100)
print(execution_time)

Generally, the number of feature points relative to the total number of images is small, so I believe this test is representative.

[spdfghi]

and sorry，I have deleted the repository where my previous pull request was located