Please how can I separate one Test Data for both Training and Testing in Machine Learning?

Hello. I am currently working on Retinal Blood Vessel Classification. I have the CenterPointProfile Data File (Which the center point X and Y of each segment) for each Image - for 20 Test Images. This File also contains the Width and the direction to which the blood vessel segment. I have read this file, Calculated the mean, median and std for the intensities for each of the segment, from each of the Colour Channels using the labelled Image. So, I have 10 features. I have saved this feature as an excel file (.xls). I put all the segments from all the 20 Test Data Images to on table with 10 columns containing the feature. I imported this into the Classification Learner App and trained this Data. I tried three different Classifier for this Data but keep getting an accuracy below 60%. I don't know how to test the system, since and I working with a single Data called the Test Data. Please someone should help me check what is wrong. Maybe my code has some wrong calculations. Please resolve this for me. Thanks

[Nziadam Nziadam]

This is dropbox link to the full Folder.

The Labelled Images The Result Folder which contains the CPprofile The DRIVE Datasets The .xls file containing all segments with labels Called Allfeatures

https://www.dropbox.com/s/34756vx2yec5r8i/My%20Project%202016%20clone.zip?dl=0