0

I am calculating sharpness scores of images (from same fov) using this algorithm:

def calc_sharpness_score(image): 
image=gaussian_filter(image,sigma=3)
kernel_size = 7 
scale = 1 
delta = 0 
lplImage = cv2.Laplacian(image, cv2.CV_64F, ksize=kernel_size, scale=scale, delta=delta) 
(mean, std) = cv2.meanStdDev(lplImage) 
return std[0][0]**2

So, how can I calculate the upper limit of this score? Or is there any upper limit for it?

Sum
  • 11
  • 2
  • how? by considering the steps of the calculation and the possible value ranges. give it a try. – Christoph Rackwitz Jan 21 '22 at 15:32
  • for minimum, I can run code on a black image and the standard deviation is 0. but how to define the sharpest image to feed the code and find the max standart deviation. it's like there may be always sharper images than the images in my list. for example, if the fov consists of circles, is the sharpest image the one consisting of perfect circles? – Sum Jan 21 '22 at 16:58

0 Answers0