----------------------------------------------README---------------------------------------
How to use the code to reproduce the results in the paper:

'thom66' is the implementation of TS-BSSCD.
'thom4' is the implementation of TS-BSSCD(O).

In this directory, the simulated data can be a vector or an image, corresponding to 'GenerateVector_vb_findcl', 'GenerateVector_vb_incontrol', 'GenerateVector_vb', 'GenerateImage_vb_findcl', 'GenerateImage_vb_incontrol', 'GenerateImage_vb', respectively. 

Take vector type as an example, to set a controlimit, you can run 'GenerateVector_vb_findcl' to get a controlimit to ensure arl=200. Or simply, you can set a controlimit in the 'GenerateVector_vb_incontrol'. Tune up or down your controlimit until the arl=200. Then you can set this controlimit in 'GenerateVector_vb' and run, to obtain the ADDs as well as STDs for a specific setting. Monitoring the image data is the same as that.

'data.mat' is the data file used in our case study. You can use 'solar_incontrolmy1' to set the controlimit to ensure the arl=1100. And run 'solar_vbmy1' to get the runlength and the test statistic. Also, you can visualize the results as the figures show in our paper.

Note that the results you obtain by running those code may look a little different from the ones in the paper because the simulated data is randomly generated. But that does not impact the strength of our method.

----------------------------------------------------------------------------------------------------