1

Let's say I have 30 of those pictures

https://i.stack.imgur.com/jzxkv.jpg

I have the numbers (amounts of dots on both sides, e.g. 17 and 19) at hand, alternatively I could also do something with the distances "red dots <-> black straight line".

I would like to quantify, that on "average", there are equal amounts of red dots on both sides of the black straight line, or in other words, the big green dot has "no tendency to be sometimes at the outer rim of the red dot cloud".

Gertrude Mauzikowski
  • 11
  • Are you given the black line initially? Computation of the distance of the black line is basically the maximum margin classifier which can be computed with SVM. If you have the counts then your problem is seemingly estimating a bernoulli distribution? – bearrito May 04 '15 at 20:43
  • The black line is a) always 90° to a tangent of the curved black line (which is given) and b) cutting through the big green dot. Could you explain the rest of your post like I am five? – Gertrude Mauzikowski May 04 '15 at 23:53
  • 1
    Im still confused about why this is tagged images processing? Do you have you to compute the location of the green dot or the radius/center of the circle from the image? I guess what in getting at is are you trying to extract features from the image (counts of dots on each side) or are you trying to quantify the discrepancy in left/right counts? – bearrito May 04 '15 at 23:59

0 Answers0