Comparative study of C-V active contour model and subdivision for micro algae image segmentation

Mass of micro algae propagated may produce harmful effects on marine ecological environment and fishery resources. The rapid and accurate identification and classification for micro algae is one of the important research issues in fisheries resource. In this paper, the comparative study of two segmentation methods, which are C-V active contour model based on partial differential equation and the 4-point approximating subdivision scheme widely used in computer aided geometric design, is given. The edges extracted by C-V active contour model can express the contour of boundary accurately but de-noise ability is not strong. The edges gained by subdivision can express the smoothness of the total contour better and the multi-resolution representation of the edges is simple but the partial accuracy is destroyed. Experiments demonstrate the validity of the comparative research between the two segmentation methods. C-V active contour model is fit for the images which have close convex contours and comparative clear boundaries. The edge extraction method based on subdivision is applicable for operations which concern total edges and ignore relatively unimportant detail information. The two methods all possess efficiency for identification and classification for micro algae images in certain conditions.