So I'm trying to find fingers using convexityDefects(); however, I am getting a lot of false positives. I thought I could fix this by going to the given convexity defect endpoint and within the contour near that area I could integrate to find the area and if it is small enough that could be determined an actual finger and not a false positive. How would I go about doing so, any help would be great Thanks!
Asked
Active
Viewed 918 times
0
-
If you solved this problem yourself, it would be very helpful if you could post your solution or a description of it. – cjm2671 Sep 03 '14 at 15:48
1 Answers
0
Well:
defects = cv2.convexityDefects(cnt,hull)
gives you a list of defects that are of the form:
[start_point,end_point,far_point,distance]
Using that you should be able to calculate the area of the triangles. However you will still likely see a number of false positives.
I've been asking a similar question here: scipy signal find_peaks_cwt not finding the peaks accurately?
