0
$\begingroup$
Why asking image processing in here?

My question is, about image processing. I can't find stackexchange site to ask about image processing, so I asking in here. Image processing is closed at computer vision, in Deeplearning.

Question

I see the SIFT algorithm, and I have some questions. The question is, "If given images are simple colored images, then SIFT algorithm can catch the keypoints?"

So, I make simple images like below with mspaint.exe enter image description here

I Think SIFT algorithm catch keypoints near edges or outsides of rectangle. But the results are quite different.

enter image description here

Please see the red rectangle at the very bottom of the image. As we can see, SIFT algorithm catch the center of the rectangle.

I think center of the rectangle has not any light changes, so I don't know why this points are keypoints.

Running code
import cv2 as cv

img=cv.imread('test.png') # read image
gray=cv.cvtColor(img,cv.COLOR_BGR2GRAY)

sift=cv.SIFT_create() 
kp,des=sift.detectAndCompute(gray,None)

gray=cv.drawKeypoints(gray,kp,None,flags=cv.DRAW_MATCHES_FLAGS_DRAW_RICH_KEYPOINTS)
cv.imshow('sift', gray)

#print(kp)
print(len(kp))
k=cv.waitKey()
cv.destroyAllWindows()
$\endgroup$

0

You must log in to answer this question.

Browse other questions tagged .