Prepare svm datasets for multi svm

Prepare Training set, Test set and Train groups(labels) for multisvm

please note that this function will give test set and training set images separately as a output and will convert images to grayscale.
This function will take four input variables and will give three output variables, which are as follows:
Input : TrainingSet_location, No_Testset,width, height
Output : Train_Set,Test_Set,train_label

Input:
TrainingSet_location : Only o/p of function DirRead is accepted or an array of locations
No_Testset : No of testset images per class.
width, height are resize parameters of images.