MATLAB: Size differ when processing

matrix manipulation

In my code below can u tell why the number of elements are different or size is different

>> img=imread('peppers.png');
>> flatImg = double(reshape(img,size(img,1)*size(img,2),size(img,3)));
idx = kmeans(flatImg,3)
size(flatImg)
196608           3
size(idx)
196608           1

please provide assistance

Best Answer

You have 196608 pixels. In flatImg you rearranged them into rows where each column is the red, green, and blue values.

With kmeans, each of the 196608 pixels is classified as a "1", "2", or a "3" depending on its color, so it's a 1D array. There is no information along any second dimension like there was with flatImg.

There is no reason why they should be the same size because they represent different things. Do you think they should?

Related Solutions

MATLAB: How to read image from pixel values

Your title does not appear to be asking the same thing as your question ??

Are you reading 10 images all exactly the same size, and you want to get a per-pixel mean across the 10 images? If so then provided that the images are gray-scale, use

mean(cat(3,IMG1,IMG2,IMG3,IMG4,IMG5,IMG6,IMG7,IMG8,IMG9,IMG10),3)

If the images are pseudocolor, then you need to use the associated colormap and ind2rgb() to convert them to RGB before you can find the mean.

If you are working with truecolor (RGB -- separate red, green, and blue pixel values), then you need to decide whether you want the mean intensity (brightness) or if you want a mean per color channel. If you want a mean intensity, use rgb2gray() and then the mean() that I showed earlier. If you want a mean per color channel, then

mean(cat(4,IMG1,IMG2,IMG3,IMG4,IMG5,IMG6,IMG7,IMG8,IMG9,IMG10),3)

(I think.)

Warning: the process of taking the mean() will leave you with a double-precision array, even if your input arrays were uint8. If your input arrays were not already double-precision, then you will not be able to directly display or imwrite() the output. You will probably want to cast() the output of the mean to class(IMG1), as in

cast(mean(cat(3, IMG1, IMG2, IMG3, IMG4, IMG5, IMG6, IMG7, IMG8, IMG9, IMG10),3), class(IMG1))

MATLAB: K-means on a gray scale image

1. kmeans only supports double or single inputs. It works in floating point so integers wouldn't make sense.

doc reshape

will answer this. It does pretty much exactly what it says - it reshapes your output back to the number of rows and columns of your image so that you have an image back rather than just a vector of class indices.

doc imshow

describes the different scaling you can do with imshow. This syntax is the 3rd in the list.

Best Answer

Related Solutions

MATLAB: How to read image from pixel values

MATLAB: K-means on a gray scale image

Related Question