Question

蓝白检测的算法

**for** each pixel in extracted region do
      **if** R > 90 and R > B and R > G then
        Mark the pixel as **healthy skin**.
      **else**
        Ignore the pixel and continue.
    **end if
    end for**
    Set R¯s as the mean of red channel values for pixels marked
    healthy skin.
    **for** each pixel in the image **do**
      nB = B/R+G+B
      rR = R/R¯s
        **if** nB ≥ 0.3 and −194 ≤ rR < −51 then
          Classify pixel as veil
        **else**
          Classify pixel as non-veil
    **end if
    end for**

我已经执行了算法的第一部分，即将像素分类为健康皮肤或非健康皮肤。如何与第二部分一起进行？

clc;
clear all; 
colorSkin=imread('veil.jpg'); %original image
% colorSkin=imresize(a, [384 512]);
[m, n]=size(colorSkin);
 hs = colorSkin; %initialising heaklthy skin array
 nhs = colorSkin; %initialising non-healthy skin array
 R = colorSkin(:, :, 1);
 G = colorSkin(:, :, 2);
 B = colorSkin(:, :, 3);
 ROI = R > 95 & R > B & R > G;
 ROI3 = ROI(:,:,[1 1 1]);
 hs(~ROI3) = 0;
 nhs(ROI3) = 0;

当我尝试使用以下命令实现算法的第二部分时：

> nB=B(hs)./(R(hs)+B(hs)+G(hs));

我收到的错误是“下标索引必须是正整数或逻辑。”

Answer 1

在你的算法中，你需要RBG 所有数据，而不仅仅是健康的皮肤，至少你写它的方式肯定是这样。否则，rR没有任何意义。另外要意识到你的伪代码必须是错误的，因为rR不可能是否定的。

Continuing from my previous answer

R=colorSkin(:, :, 1);
G=colorSkin(:, :, 2);
B=colorSkin(:, :, 3);
skin=repmat(R>90 & R>B & R>G,1,1,3);

hs=colorSkin;
hs(~skin)=0;
nhs(skin)=0;

% new code
nB = B./(R+G+B); % I assume you dont mean B./R+G+B as in the pseudocode....
rR=R./mean(R(skin(:,:,1)>0);

% use the same code as before, changing the indexes to classify veil

如何在MATLAB中为以下算法编写if-else条件语句？

1 个答案: