Ask Your Question

Detect elipse/oval shape in Black and White image

asked 2014-01-22 09:01:52 -0600

2hochn gravatar image

I want to detect dark foot-shapes in a picture. Shapes on noisy background

For this i want to start of by using a treshhold and search the resulting picture for shapes of the correct size.

After the Tresholding, the dark shapes are (most of the time) pretty good recognizable. My Question is: What is the best/fastest way to detect the shapes? Related Question

A few remarks: First i used findContours and used contourArea to get the size of the shapes. This works fairly well. I thought that it might speed things up to use Connected Component Labeling while simultaneously computing the size of the regions i find. After i was done computing this, i found the paper "A linear-time component-labeling algorithm using contour tracing technique", that claims, that it's much faster to compute only the contours instead of the CC-labeling. Now I'm wondering, what Algorithm findContours in OpenCV uses and what algorithm i should use.

edit retag flag offensive close merge delete



If you actually look at the documentation, you can see which algorithm the implementation uses, the paper is even referenced! Actually I am not sure there is anything faster then just using the approach you suggest, segmentation + contour finding is already realtime right? Why should it go faster?

StevenPuttemans gravatar imageStevenPuttemans ( 2014-01-23 04:49:01 -0600 )edit

1 answer

Sort by ยป oldest newest most voted

answered 2014-01-23 08:52:15 -0600

NikolasE gravatar image

How does your thresholded image look like? Have you tried to simply downsample it?

edit flag offensive delete link more

Question Tools


Asked: 2014-01-22 09:01:52 -0600

Seen: 1,087 times

Last updated: Jan 23 '14