Re: hough ellipse fit inaccurate?
Ok sorry, maybe I need to explain my project better. I have a Autodesk Maya Model of a simplified eyeball with the pupil (see image). I render the images of the eyeball from different angles and try to detect the pupil center as accurate as possible. Since I know the geometry http://www.dict.cc/englischdeutsch/geometry.html and rotation of the eyeball, I can calculate the mapping of the real center of the pupil on my virtual maya camera sensor. I proved the validity http://www.dict.cc/englischdeutsch/validity.html of my calculation with several other methods of ellipse center detection (center of mass, distance transform, opencv ellipse fit, starbust) where I get errors between my calculation and measurement with sometimes less than 0.02 pixels. But nevertheless http://www.dict.cc/englischdeutsch/nevertheless.html I want to try the hough ellipse approach, because it may be more robust against noise or other errors I want to simulate later. And it seems so far only the hough ellipse approach is quite inaccurate so I was wondering why. I think the reason could be the ellipse center detection has only a half pixel accuracy while my other approaches have subpixel accuracy. I think posting the calculation code would be too much, but I am quite sure the calculation is right. Kind regards Arno
Hi Arno, Looking at the code, I would ask: Did your score improve by setting accuracy=1 ? https://github.com/scikitimage/scikitimage/blob/master/skimage/transform/_... Considering that you are asking for accuracy below half a pixel, I would not be surprised if the voting process of the Hough transform is not that accurate. A leastsquare fitting (Opencv fitellipse) might be more accurate than a voting process for a perfect ellipse. Aren't the eyeball and the pupil both balls? If you slice them in any way, wouldn't you obtain disks? So why detecting elllipses and not circles? Maybe hough_circle will be more accurate. Sorry I cannot provide any proof or certitude on how accurate hough_ellipse is. Kind regards, Kevin
Hi Arno, Using the image you provided here, I fit the black pupil center using a ellipse estimation algorithm. The result is 295.3148 along the horizontal direction from left to right 106.7683 along the vertical direction from the top to down. Since I don't know the exact pupil center for this image, I‘m not sure the result is good or not. Best wishes! Xu Cheng 在 2015年3月5日星期四 UTC+8下午7:09:34，Arno Dietz写道：
Ok sorry, maybe I need to explain my project better. I have a Autodesk Maya Model of a simplified eyeball with the pupil (see image). I render the images of the eyeball from different angles and try to detect the pupil center as accurate as possible. Since I know the geometry http://www.dict.cc/englischdeutsch/geometry.html and rotation of the eyeball, I can calculate the mapping of the real center of the pupil on my virtual maya camera sensor. I proved the validity http://www.dict.cc/englischdeutsch/validity.html of my calculation with several other methods of ellipse center detection (center of mass, distance transform, opencv ellipse fit, starbust) where I get errors between my calculation and measurement with sometimes less than 0.02 pixels. But nevertheless http://www.dict.cc/englischdeutsch/nevertheless.html I want to try the hough ellipse approach, because it may be more robust against noise or other errors I want to simulate later. And it seems so far only the hough ellipse approach is quite inaccurate so I was wondering why. I think the reason could be the ellipse center detection has only a half pixel accuracy while my other approaches have subpixel accuracy.
I think posting the calculation code would be too much, but I am quite sure the calculation is right.
Kind regards Arno
participants (3)

Arno Dietz

Kevin Keraudren

xu cheng