X. Bi and X. Sun (PRC)
Ant colony algorithm, clustering, texture feature, image recognition
In this paper a new texture image recognizing method was proposed based on the characteristics of ant colony algorithm such as characteristic of clustering. The effective feature function values of the texture images are proposed from the aspect of the characteristic of texture images. And these texture feature function values can be clustered, which can recognize many texture images. The experimental results show that the new method in this paper can improve the recognizing rate of image effectively. It is an effective method in image recognizing. 2.1 Ant colony algorithm The first ant colony optimization called ant system was inspired through studying of the behavior of ants in 1991. These methods solved TSP (traveling salesman problem) such as assignment problem, job-shop scheduling problem (JSP) and so forth earliest, and received better effect. Now for the sake of theory of ACS, take TSP as example to describe the algorithm. Suppose a directive graph is defined as G=, where N is a set of cities, and E is a set connecting edges of two cities. Define a set of paths is , feasible solutions of TSP, which connecting an initiative city and a terminal city through a series of interim city by edges. Ants are placed in the cities arbitrarily. In this problem, every ant will choose a path 1 to N if there are N cities. Define ( , is an edge of city i and city j. ω )i j
Important Links:
Go Back
