文档下载 文档下载导航
设为首页 | 加入收藏
搜索 请输入内容:  
 导航当前位置: 文档下载 > 所有分类 > TWO-DIMENSIONAL ENTROPY METHOD BASED ON GENETIC ALGORITHM
侵权投诉

TWO-DIMENSIONAL ENTROPY METHOD BASED ON GENETIC ALGORITHM

Abstract-Image segmentation represents the first step in many applications in image understanding. Thresholding is a simple and important in image segmentation. One-dimensional entropy does not take into account the spatial correlation between the pixels i

3) Replace each individual’s fitness by its new fitness which is caculated as

N

fiti

Ni ∑fiti

i=1

N

.

Then, get the left Nreproduction.

∑ Ni individuals for the next offspring by applying the simple proportional

i=1

By combining random selection techniques and certain selection techniques, RSSR reproduction assures individuals with higher fitness than average fitness being selected for recombination, thus turns to have less selection error.

3.2 Two-point crossover and simple mutation

One-point crossover employed in SGA chooses a point at random, called the crossover point, and exchanges the segments to the right of this point. It has one major drawback in that certain combinations of schema cannot be combined in some situations.

Two-point crossover chooses two points at random, and exchanges the segments between them. Research indicates that two-point crossover performs better than one-point crossover[10].

Simple mutation [8] is related to a bit (gene) that could be changed according to the mutation probability. 3.3 The elitist strategy [8]

Although the RSSR reproduction can protect the best individual of each generation, it is still possible for the best individual to be disrupted by the crossover and mutation. Thus it is necessary to introduce the elitist strategy by which the best individual is copied into the succeeding generation directly.

4. Image segmentation based on Genetic Algorithm using two-dimensional entropy

There are two major considerations in practical application of GA: one is how to map the optimization problem into GA’s search space, i.e., how to encode the individuals so as to solve the optimization problem. The other one is how to choose proper fitness function.

The image segmentation deals with pixels of different gray-level. Since the practical image used in this paper is 128-level, each gray level in the population is represented by 7 bits. Since 2-dimensional histogram contain both the pixel gray level and the average gray level of the neighborhood, the chromosomes are encoded as 14 bit strings, with the first 7 bits representing the pixel gray level and the second 7 bits representing the average gray level of the neighborhood.

The fitness function is derived from two-dimensional entroopic method, which is defined as

ψ(s,t)=H(O)+H(B)=lnPst(1 Pst)+Hst+(HL 1L 1 Hst1 P).

stst

In this approach, the population size is 10, the crossover probability is 0.8, the mutation probability is 0.01, predetermined number of generations is 40.

An abstract procedure of the Improved Adaptive Genetic Algorithms is given below, where population of candidate solutions to a given problem at generation

(1)

N(k) is a

k:

k=0,generate an initial population N(k).

N(k).

(2) Compute the population

(3) Perform the GA’s reproduction, crossover, and mutation.

(4) Finish the GA and get the final optimal threshold if the predetermined number of generations is reached or

上一页第5页

你可能喜欢

热门文档

相关文档

站点地图 | 文档上传 | 侵权投诉 | 手机版
本站所有资源均来自互联网,本站只负责收集和整理,均不承担任何法律责任,如有侵权等其它行为请联系我们.
文档下载 Copyright 2013 doc.guandang.net All Rights Reserved.  -  email