notes.txt

#patient_tags = ds.dir("patient")
    #for x in patient_tags:
    #    print(ds.data_element(x))
    #print(ds)

    #image_data = ds.pixel_array
    #print("Image length = ", len(image_data))
    #print("Rows = ", ds.Rows)
    #print("Rows = ", ds.Rows)
    #print("Image length = ", len(image_data))
    #image = Image.frombytes('RGBA', (256,256), image_data)
    #image.show()

    #image_arr = ds.pixel_array
    #im = Image.fromarray(cm.gist_yarg(image_arr)).show() //bad
    #plt.imshow(ds.pixel_array, cmap='gray')    /good
    #plt.show()


from msilib.schema import File

from write_np_to_file import write_nparr_to_file


we have ascertained that the pixel data remains consistent (changes after watermarking and is restored correctly)
and so does the metadata (does not change). For this we printed the metadata to a file at various points and hashed the contents of the file
using Get-FileHash in powershell. For the numpy array we used the function in write_nparr_to_file to dump the whole 
array to a file and compared hashes again at various points. 
Thus we found they were consistent. Proof is in this image

to find time per module do
import time
st = time.time()

and at the end time.time()-st