image processing, dimension reduction

2018 3 19
1 个回答
回答已采纳
更新时间：2018 3 20
2 次查看（30 天）

0 个投票

if i have number of images (multi spectral images for the same scene), and i want to reduce the dimensional of this images by any dimension reduction techniques where in the result i get one reducing image for 200 as example. how can i do this task? thank you.

请先登录，再回答此问题。

 采纳的回答

Image Analyst
Image Analyst 2018-3-19

1 个投票

See my attached demo for pca. It should be easy for you to adapt it to take any number of spectral bands instead of the 3 visible color (RGB) bands I took. If you want just one single image that represents most of the variation, take the principal components #1 image.

4 个评论
显示 2更早的评论 隐藏 2更早的评论

i try your demo sir, and it's work, but if i have more than 9 images how can i applying this demo for it?
Adam
Adam 2018-3-19
A demo is just an example of how the code works, you usually need to adapt the code for your own purposes.
Image Analyst
Image Analyst 2018-3-19
Two code samples for processing multiple files in a folder are shown in the FAQ : http://matlab.wikia.com/wiki/FAQ#How_can_I_process_a_sequence_of_files.3F
ok, thanks all; another question please, i wrote an code for (KLT) techniques for dimension reduction for image (I'm just a beginner in matlab my code maybe is not good enough one or it's want some modification) the result for the code is binary image, is it possible that i reduce the dimensionality without losing the color (can i get colored reducing image as result)
clc, clear all, close all x = imread ('10.bmp'); whos %x2 = imread ('2.bmp'); %x3 = imread ('3.bmp'); %x4 = imread ('4.bmp'); %x5 = imread ('5.bmp'); %x6 = imread ('6.bmp'); %x=[x1 x2 x3 x4 x5 x6]; t=graythresh(x) z=im2bw(x,t); [n m]=size(z);
for j=1:m mu(:,j)=mean(z(:,j)); end for j=1:m xx(:,j)=z(:,j)-mu(j); end xx=xx'; cx=(xx*xx')./(n-1); [M V]=eig(cx);
%my
V; size(V) mm=max(max(V)) ee=sum(diag(V));
%
A=(rot90(M))'; y=A*xx; Ay=A'*y; for j=1:m xhad=Ay+mu(j); end
k=input('k='); for i=1:k for j=1:k xxk(i,j)=xx(i,j); cxk(i,j)=cx(i,j); end end for j=1:k muk(:,k)=mean(xxk(:,k)); end
[M1 V1]=eig(cxk); Ak=(rot90(M1))'; yk=Ak*xxk; Ayk=Ak'*yk; for j=1:k xhadk=Ayk+mu(j); end
figure; imshow(xhadk),title('A=510')
figure; subplot(3,2,1),imshow(x); subplot(3,2,2),imshow(xhad); subplot(3,2,3),imshow(xhadk);

请先登录，再进行评论。

更多回答（0 个）

类别

在 帮助中心 和 File Exchange 中查找有关 Dimensionality Reduction and Feature Extraction 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