【图像压缩】一种基于香农熵和差分进化算法的多级图像阈值的图像压缩新方法附matlab代码

1 内容介绍

图像压缩是图像传输和存储中最重要的步骤之一。 大多数最先进的图像压缩技术都是基于空间的。 在此代码中，基于多级图像阈值实现了基于直方图的图像压缩技术。 图像的灰度被划分为一组清晰的概率分区。 香农熵用于衡量清晰分组的随机性。 熵函数使用一种流行的称为差分进化的元启发式算法来最大化，以减少优化目标值的计算时间和标准偏差。​

2 仿真代码

function entropy=shannonEntropy(x,h)

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

% INPUTS : x is the threshold vector

%          h is the histrogram of the image

% OUTPUT:  entropy is the Entropy of the image after probabilistic 

%          partition by the threshold vector 'h'

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

% Written by : Sujoy Paul, Jadavpur University, Kolkata %

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

entropy=0;

x=[1 (x+1) 256];

s=size(x);

for i=1:s(2)-1

    temp = h(x(i):x(i+1));

    tsum= sum(temp);

    ent=0;

    if tsum~=0

    for j=x(i):x(i+1)

        if h(j)~=0

            a=h(j)/tsum;

            ent = ent + a*log(a);

        end

    end

    end

    entropy = entropy + ent;    

end

entropy=-entropy;

3 运行结果

4 参考文献

[1] Paul S . A novel approach for image compression based on multi-level image thresholding using Shannon Entropy[J]. Differential Evolution.

博主简介：擅长智能优化算法、神经网络预测、信号处理、元胞自动机、图像处理、路径规划、无人机等多种领域的Matlab仿真，相关matlab代码问题可私信交流。

部分理论引用网络文献，若有侵权联系博主删除。