PyDWI 0.0.1

Fast ADC and DKI maps

PyDWI

Fast ADC and DKI maps from DWI.

Install

pip install pydwi

How to use

PyDWI supports interactive usage only at the moment. Command-line usage is planned for the future.

from PyDWI.core import DWIDataset, get_ADC_dataset, get_DKI_dataset, save_nii, show

First, instantiate a DWIDataset class:

data = DWIDataset("1.dcm")

Loading and rescaling...
Successfully loaded Dataset

You can have get information about the dataset by printing data.info

print(data.info)

DICOMDataset with 185 slices in groups of 5 slices each
        rescaled with slope 1.7836 and intercept 0.0000.

You can get information about the logical slice groups using data.slice_groups

data.slice_groups

[SliceGroup at position [-83.12] with instance numbers [4 8 12 16 20],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [-80.12] with instance numbers [24 28 32 36 40],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [-77.12] with instance numbers [44 48 52 56 60],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [-74.12] with instance numbers [64 68 72 76 80],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [-71.12] with instance numbers [84 88 92 96 100],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [-68.12] with instance numbers [104 108 112 116 120],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [-65.12] with instance numbers [124 128 132 136 140],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [-62.12] with instance numbers [144 148 152 156 160],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [-59.12] with instance numbers [164 168 172 176 180],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [-56.12] with instance numbers [184 188 192 196 200],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [-53.12] with instance numbers [204 208 212 216 220],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [-50.12] with instance numbers [224 228 232 236 240],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [-47.12] with instance numbers [244 248 252 256 260],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [-44.12] with instance numbers [264 268 272 276 280],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [-41.12] with instance numbers [284 288 292 296 300],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [-38.12] with instance numbers [304 308 312 316 320],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [-35.12] with instance numbers [324 328 332 336 340],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [-32.12] with instance numbers [344 348 352 356 360],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [-29.12] with instance numbers [364 368 372 376 380],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [-26.12] with instance numbers [384 388 392 396 400],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [-23.12] with instance numbers [404 408 412 416 420],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [-20.12] with instance numbers [424 428 432 436 440],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [-17.12] with instance numbers [444 448 452 456 460],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [-14.12] with instance numbers [464 468 472 476 480],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [-11.12] with instance numbers [484 488 492 496 500],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [-8.12] with instance numbers [504 508 512 516 520],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [-5.12] with instance numbers [524 528 532 536 540],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [-2.12] with instance numbers [544 548 552 556 560],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [0.88] with instance numbers [564 568 572 576 580],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [3.88] with instance numbers [584 588 592 596 600],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [6.88] with instance numbers [604 608 612 616 620],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [9.88] with instance numbers [624 628 632 636 640],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [12.88] with instance numbers [644 648 652 656 660],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [15.88] with instance numbers [664 668 672 676 680],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [18.88] with instance numbers [684 688 692 696 700],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [21.88] with instance numbers [704 708 712 716 720],
                  b_values [0 50 300 600 1000] and 5 slices.,
 SliceGroup at position [24.88] with instance numbers [724 728 732 736 740],
                  b_values [0 50 300 600 1000] and 5 slices.]

len(slice_groups) gives you the number of usable slices

len(data.slice_groups)

37

You can get additional information on the slice groups:

data.slice_groups[0]

SliceGroup at position [-83.12] with instance numbers [4 8 12 16 20],
                 b_values [0 50 300 600 1000] and 5 slices.

You can see some slices by calling show:

data.slice_groups[0].show()

You can get ADC maps by calling get_ADC_dataset. By default this runs in parallel and is quite fast.

%%time
ADC_maps = get_ADC_dataset(data)

CPU times: user 2.32 s, sys: 42.3 ms, total: 2.36 s
Wall time: 310 ms

You can get DKI maps by calling get_DKI_dataset. This also runs in parallel but takes around 10 minutes per dataset for around 30 slices.

%%time
D_maps, K_maps = get_DKI_dataset(data)

[Parallel(n_jobs=-1)]: Using backend LokyBackend with 16 concurrent workers.


CPU times: user 1.11 s, sys: 304 ms, total: 1.41 s
Wall time: 10min 1s


[Parallel(n_jobs=-1)]: Done  37 out of  37 | elapsed: 10.0min finished

You can have a look at slices by calling show which is a utility function

show(D_maps[0])

<matplotlib.image.AxesImage at 0x7fa5c99021d0>

show(K_maps[0])

<matplotlib.image.AxesImage at 0x7fa5bb4d4f90>

Finally, you can save .nii.gz files by calling save_nii and providing a filename without the nii.gz.

save_nii(D_maps, "D_maps")