Query about the kNN and SVM classifiers training

Hi all,

could anyone help me how to use the kNN and SVM classifiers where my problem in details as follows :

1-I have two classes of fruit images (Normal and Abnormal)

2-I extracted color feature histogram where bins=250

3-I sorted these extracted features in a matrix of size 250 *100

4-I transposed each matrix like : transposed_matrix= matrix' % for K-means clustering purpose

5-I clustered each matrix to 4-centers using K-means algorithm where each center is a feature

6- I stored 40 clustered matrices of normal images (each one is represented by 4-centers ) in the training matrix to be 160 centers * 250 columns as a training matrix ,also I did the same thing for abnormal images in another training matrix also

But now I don't know how to give these two training matrices to the kNN and SVM classifiers

and then test an unknown image taken from the testing set (also represented by 4-centers) using

kNN and SVM classifiers to classify it to normal or abnormal class,could anyone guide me how to

do that in a matlab code lines because I am new in matlab I saw the svmclassify function on

http://www.mathworks.com/help/stats/svmclassify.html

and

http://www.mathworks.com/help/stats/svmtrain.html

But I don't know how to adapt these codes for my case, I attached my test code

 and thank you so much in  advance.