Add noise to image-datastore using "Randn" command on Matlab

2 次查看(过去 30 天)
I have used command Randn, but still give me the same accuraccy result even after changing the noise magnitude. Why accuracy still same? how image can be affected by noise
used commands :
A=imread('human_6_77_.png');
K = imcrop(A,[116 50 675 534]);
b = 20*randn(size(K));
I=uint8(double(K)+b);
imshow(I)
  2 个评论
Sai Sri Pathuri
Sai Sri Pathuri 2020-7-11
Here are the images I tried with your code, i see a change in both images. There is more distortion for image with noise magnitude of 100. Can you attach your images? What did you mean by accuray? Is it the same as amount of distortion or something else?
% Original Image
figure
A=imread('coins.png');
K = imcrop(A,[116 50 675 534]);
I=uint8(double(K));
imshow(I)
% Image with Noise magnitude = 20
figure;
A=imread('coins.png');
K = imcrop(A,[116 50 675 534]);
b = 20*randn(size(K));
I=uint8(double(K)+b);
imshow(I)
% Image with Noise magnitude = 100
figure;
A=imread('coins.png');
K = imcrop(A,[116 50 675 534]);
b = 100*randn(size(K));
I=uint8(double(K)+b);
imshow(I)
Mohamed Elbeialy
Mohamed Elbeialy 2020-7-14
As changing magnitude, the noise will increase on image. However, when trying to train this image on matlab "deep learning", it keeps givening me the same training accuracy evern with different noise magnitudes. why? what idology is running ?

请先登录,再进行评论。

回答(0 个)

类别

Help CenterFile Exchange 中查找有关 Image Data Workflows 的更多信息

产品


版本

R2019a

Community Treasure Hunt

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

Start Hunting!

Translated by