matlab人脸计数

基于引用[1]中的肤色定理和形态学方法，可以使用MATLAB进行人脸计数。具体步骤如下：

1. 读取图像并转换为YCbCr颜色空间。

img = imread('face.jpg');
img_ycbcr = rgb2ycbcr(img);

2. 根据肤色定理，提取肤色区域。

Cb = img_ycbcr(:,:,2);
Cr = img_ycbcr(:,:,3);
skin = zeros(size(img_ycbcr,1),size(img_ycbcr,2));
for i = 1:size(img_ycbcr,1)
    for j = 1:size(img_ycbcr,2)
        if Cr(i,j) >= 133 && Cr(i,j) <= 173 && Cb(i,j) >= 77 && Cb(i,j) <= 127
            skin(i,j) = 1;
        end
    end
end

3. 对肤色区域进行形态学处理，去除小面积干扰。

se = strel('disk',5);
skin_close = imclose(skin,se);
skin_open = imopen(skin_close,se);
skin_final = bwareaopen(skin_open,1000);

4. 对肤色区域进行连通区域分析，得到人脸个数。

[L,num] = bwlabel(skin_final,8);

完整代码如下：

img = imread('face.jpg');
img_ycbcr = rgb2ycbcr(img);

Cb = img_ycbcr(:,:,2);
Cr = img_ycbcr(:,:,3);
skin = zeros(size(img_ycbcr,1),size(img_ycbcr,2));
for i = 1:size(img_ycbcr,1)
    for j = 1:size(img_ycbcr,2)
        if Cr(i,j) >= 133 && Cr(i,j) <= 173 && Cb(i,j) >= 77 && Cb(i,j) <= 127
            skin(i,j) = 1;
        end
    end
end

se = strel('disk',5);
skin_close = imclose(skin,se);
skin_open = imopen(skin_close,se);
skin_final = bwareaopen(skin_open,1000);

[L,num] = bwlabel(skin_final,8);
disp(['人脸个数为：',num2str(num)]);