0 个投票

I have 3 images that I have to label using the knn algorithm and another 18 images that i know. I don't know how to label those 18 img using a variable 18x1 named etichete and don't know how to classificate them using the knn. ext_auto is the function for maximum contrast extension. And it doesn't show hist1,hist2, hist3.
  • maximum contrast extension, adaptive variant
  • histogram calculation
  • patient identification using the algorithm for classifying the closest k-neighbors (k-nearest neighbor - knn), the comparison will be made at the histogram level.
img1 = imread('img34.png');
img2 = imread('img35.png');
img3 = imread('img36.png');
img1 = ext_auto(img1,1);
img2 = ext_auto(img2,1);
img3 = ext_auto(img3,1);
hist1 = histogram(img1,'Normalization','probability');
hist2 = histogram(img2,'Normalization','probability');
hist3 = histogram(img3,'Normalization','probability');
figure,
subplot(2,3,1), imshow(hist1);
subplot(2,3,2), imshow(hist2);
subplot(2,3,3), imshow(hist3);
locatie = 'C:\Users\User\Desktop\Facultate\Anul III\Semestrul II\tsim\Tema2\antrenare';
files=dir(fullfile(locatie, '*.png'));
load('etichete.mat');
h = zeros(length(files),256);
for i = 1:length(files)
FileName = fullfile(locatie, files(i).name);
img = imread(FileName);
img = ext_auto(img,1);
hist = histogram(img,'Normalization','probability');
figure(i), imhist(img);
end

请先登录，再回答此问题。

 采纳的回答

Image Analyst
Image Analyst 2020-5-8

1 个投票

You can't use imshow on a histogram object. You need to use plot() or bar():
counts = hist1.Values;
binCenters = (hist1.BinEdges(1:end-1) + hist1.BinEdges(2:end)) / 2;
bar(binCenters, counts);
xlabel('Gray Level', 'FontSize', 15);
ylabel('Counts', 'FontSize', 15);
Warning: If your image is RGB, the counts are of all the gray levels, regardless of what color channel they're in!

更多回答（0 个）

类别

在 帮助中心 和 File Exchange 中查找有关 Statistics and Machine Learning Toolbox 的更多信息

产品

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by