An effective multi-object C–V segmentation model is proposed.
A rapid initial contour search is conducted.
To achieve the multi-object segmentation the image is converted to a single level set.
The evolution of the curve is controlled adaptively via the contour line gradient.
The model takes the local characteristics of each target into account through the broad contours to avoid missing the target.