Chord midpoint randomized hough transform for the cell image segmentation

Randomized Hough transform is an effective approach for circle detection. However, a lot of ineffective sampling and the ineffective accumulation are introduced by the random sampling. Chord midpoint randomized Hough transform algorithm is employed in this paper for the circle detection in the cell image. So the problem of ineffective sampling can be well solved by using filter and the canny operator. At the same time, ineffective accumulation is well eliminated by employing the coordinate relationship to determine whether the circle parameters should be solved from the two sampled points. Then, according to the fact that the inscribed circles of all points on the circle must pass the circle center, all the midpoints of circle chords will be accumulated to detect the position of circle center. Experimental results indicate that the proposed algorithm performs faster and locates the cells accurately.