Content of the directory: 

Originally I used five splits but I am unable to make them available
here because of file size. You will only find one split here
(splitNo=1) please drop me an email if you also need the other ones.

train: caltech101_nTrain{nrTrainImages}_N{splitNo}_feat{featNo}.mat
test: caltech101_nTrain{nrTrainImages}_N{splitNo}_test_feat{featNo}.mat
labels: caltech101_nTrain{nrTrainImage}_N{splitNo}_labels.mat

where
	'nrTrainImages' - is the number of training images in 5,10,15,20,25,30 
	'splitNo' - number in 1,2,3,4,5 represents different splits of Caltech101. 
	'featNo' - a unique ID for an image feature (try explain_feature(featNo))

the train/test files contain a matrix in single precision of size (n train images, m test images)
    n x n for training
    m x n for testing 


the label file contains 

train labels: tr_label
test labels: te_label
train filenames: tr_files
test filenames: te_files