Method for extracting curling wheel on basis of two-stage ant colony algorithm

Abstract

The invention belongs to the technical field of iridodiagnosis, and particularly relates to a method for extracting a curling wheel on the basis of a two-stage ant colony algorithm. According to the method, secondary iteration needs to be carried out, a first stage has a task of carrying out iteration for a small number of times by using a great quantity of artificial ants to determine an approximate position of an ANW, and the position is used for determining initial positions of the artificial ants in a next stage; and in a second stage, a small quantity of artificial ants are used for carrying out repeated iteration to obtain a final curling wheel extraction result. In each iterating process, by introducing a guiding factor in a local direction and a guiding factor in a global direction, the artificial ants can be rapidly close to the ANW direction; and by using a pheromone update strategy based on a point density, concentration of pheromones near an edge of the ANW can be rapidly increased, and finally, according to a concentration value of the pheromone, a final ANW curve is extracted.

Description

technical field [0001] The invention belongs to the technical field of iris diagnosis, and in particular relates to a method for extracting crimping wheels based on a two-stage ant colony algorithm. Background technique [0002] In iris diagnostics, the iris curl wheel is an important diagnostic sign. It is located on the periphery of the intestinal ring and is a ring-shaped line that divides the iris into two parts: the pupil in the center and the ciliary body in the periphery. It is composed of blood vessels and nerves, representing the visceral autonomic nervous system (sympathetic and parasympathetic) in the human nervous system, and it can display a lot of information related to the intestinal tract and visceral nerves. When the edge of the crimping wheel is abnormal, such as expanding or contracting, it indicates that the function of the human intestinal tract is abnormal. Therefore, whether the human intestinal tract is healthy can be known by detecting the change of...

AI Technical Summary

Problems solved by technology

[0003] Due to the complex texture of the edge of the crimping wheel, traditional methods such as edge detection operator Canny operator, Prewitt operator, Roberts operator, and edge tracking cannot extract the edge of the crimping wheel well.

For the extraction of such complex edges, the literature [M.Kass, A.Witkin, D.Terzopoulous, Snakes: active contour models, in: Proceeding of the First International Conference on Computer Vision, IEEE Computer Society Press, London, 1987, pp.259-268.] proposed a method based on improved Snake The model’s crimping wheel extraction method, which overcomes the shortcomings of the traditional Snake model, such as slow convergence speed and the iterative effect being too dependent on the selection of the initial snake point position. Better, but when the edge profile of the crimping wheel is blurred or incomplete, there will be errors between the extracted crimping wheel and the real edge

In fact, the edge contour of the iris crimping wheel is often not very clear and discontinuous, so for this kind of crimping wheel, the effect of using this method to extract the crimping wheel is not very ideal

Images