how to decrease number of descriptors when using SIFT?

Question

I have used SIFT to find keypoints and descriptors,now I want to decrease number of descriptors because for example for an image with size 256x256 it gives me 120000 descriptors which is time consuming. please let me know any suggestion how to decrease number of descriptors. Thanks

This seems like too many detections for a properly implemented detector. What SIFT library are you using? What are the parameter settings you're using? What is the image that gives you this many detections? — , Nov 26 '12 at 05:14
@ Sancho: I have used [image, descrips, locs] = sift('image.tif'); I have tested in different images, for example in .tif image with the size of 256x256 — user615864, Dec 03 '12 at 05:10
Can you edit your question to include one of the example images that produces this many detections? — , Dec 03 '12 at 05:50

score 0 · Answer 1 · answered Oct 19 '13 at 16:55

0

You can’t have 120000 keypoints for an image which contains only 65536 (256x256) pixels!

answered Oct 19 '13 at 16:55

Raluca

280
2
9

score 0 · Answer 2 · answered Oct 22 '13 at 17:34

The number of key points is equal to the number of sift descriptors, unless we are using multi-scale extraction. Raluca is correct. You can't have 120 000 key points in that image but we can have 120 000 descriptors if, as I said you are extracting features at a multi-scale process. Please enlighten us with more details of the work you have done, especially what implementation you are using.

how to decrease number of descriptors when using SIFT?

2 Answers2