CTU

CT Scan Utilities. A set of useful methods for working with CT scan dicom files as simple as possible in python.

dependencies:

• pydicom
• numpy
• scipy
• matplotlib
• plotly

Read a single slice from a CT Scan:

from ctu import reader
slice = reader.ReadSlice("path to CT slice")
print(slice)

slice is a dicom DataSet object that contains pixel data and other metadata. Dataset object is the primary object and is used in many methods.

Read a CT scan

from ctu import reader
ct_slices = reader.ReadCT("path to a CT Scan directory")
print(len(ct_slices))

Plot Slices of a CT Scan:

from ctu import reader, display
slices = reader.ReadCT("path to a CT Scan directory")
display.PlotSlices(slices,wtype='lung',nrows=5,ncols=5, step=2)

wtype can be a string from ['lung' , 'bone' , 'liver' , 'tissues'] or a tuple which specifies custom WindowCenter and WindowWidth.

Working with Image data
Get image in Hounsfield scale:

from ctu import reader, display
slice = reader.ReadSlice("path to a slice")
img = reader.getImageHU(slice)
display.Plot(img)

Get image in a certain Windowing condition:

from ctu import reader, display
slice = reader.ReadSlice("path to slice")
img = reader.getWindowedImage(slice, wtype='lung')
display.Plot(img) 

Plot a CT Scan in 3D

from ctu import reader, display
slices = reader.ReadCT("path to a CT Scan folder")
display.Plot3D(slices, threshold=400)

Mask body part in CT slices

from ctu import reader, transform
slice = reader.ReadSlice("path to a slice")
img = reader.getWindowedImage(slice, 'lung')
mask, masked_image = transform.BodyMask(img) 

Lung mask of a CT Slice

from ctu import reader, transform
slice = reader.ReadSlice("path to a slice")
img = reader.getImageHU(slice)
label, mask, masked_image = transform.LungMask(img) 

K-Means clustering is used to detect different parts in image. "lable" is the output of K-Means clustering and "mask" is the computed lung mask.