Knn Classify using to classify similar images

Hello All,

I'd like to ask if someone can help me to solve a problem that i'm having with one of my project.

I'm trying to classify different images using knn as algorithm, my real goal is analyse a lung x-ray(sample) and classify sample data based on my database (training) into healthy lung, cancer lung, etc (groups) but just to simplify the understanding i'll use simple images.

I have three 100x100 pixels images to use as 'training' that are: red.jpg,green.jpg and blue.jpg

And as my sample images i have another three 100x100 pixels images that are: pink.jpg,aqua.jpg and light_green.jpg

and as a group, [1,2,3] (RED,GREEN,BLUE) to use.

Here is my code:

>> red = imread('images/red.jpg');
>> green = imread('images/green.jpg');
>> blue = imread('images/blue.jpg');
>> pink = imread('images/pink.jpg');
>> aqua = imread('images/aqua.jpg');
>> light_green = imread('images/light_green.jpg');
>> training = [red,green,blue];
>> sample = [aqua,pink,light_green];
>> group = [1,2,3];

But if i try to use the knn classify here i got this error:

Error using knnclassify (line 91)
The length of GROUP must equal the number of rows in TRAINING.

My images sizes are:

    >> size(red),size(blue),size(green),size(pink),size(aqua),size(light_green);

ans =

     100   100     3

ans =

     100   100     3

ans =

     100   100     3

ans =

     100   100     3

ans =

     100   100     3

And the sizes of parameters to knn classify are:

>> size(training),size(sample),size(group)

ans =

     100   300     3

ans =

     100   300     3

ans =

       1     3

But on this example i just have 3 groups, RED (1), GREEN (2) and BLUE (3) how can my group rows be the same as my training rows?

Thank you