jcreinhold

#/usr/bin/env python
# -*- coding: utf-8 -*-
"""
nii_to_tif

convert all mha files in a directory to nifti

Author: Jacob Reinhold (jacob.reinhold@jhu.edu)
"""

jcreinhold

import fastai.vision as faiv
import nibabel as nib

def open_nii(fn:str) -> faiv.Image:
    """ Return fastai `Image` object created from NIfTI image in file `fn`."""
    x = nib.load(str(fn)).get_data()
    return faiv.Image(torch.Tensor(x))
    
class NiftiItemList(faiv.ImageItemList):
    """ custom item list for nifti files """

jcreinhold

from functools import singledispatch

@faiv.TfmPixel
@singledispatch
def crop(x, pct, axis:int) -> torch.Tensor:
    """" crop a 3d image along an axis """
    s = x.shape
    i0, i1 = int(s[axis]*pct[0]), int(s[axis]*pct[1])
    return x[np.newaxis,i0:i1,:,:].contiguous() if axis == 0 else \
           x[np.newaxis,:,i0:i1,:].contiguous() if axis == 1 else \

jcreinhold

idb = (NiftiNiftiList.from_folder(data_dir/'t1', extensions=('.gz'))
       .split_by_folder()
       .label_from_func(get_y_fn)
       .transform((tfms,tfms), tfm_y=True)
       .databunch(bs=2))

jcreinhold

def conv3d(ni:int, nf:int, ks:int=3, stride:int=1, pad:int=1, norm='batch'):
    bias = not norm == 'batch'
    conv = faiv.init_default(nn.Conv3d(ni,nf,ks,stride,pad,bias=bias))
    conv = spectral_norm(conv) if norm == 'spectral' else \
           weight_norm(conv) if norm == 'weight' else conv
    layers = [conv]
    layers += [nn.ReLU(inplace=True)]  # use inplace due to memory constraints
    layers += [nn.BatchNorm3d(nf)] if norm == 'batch' else []
    return nn.Sequential(*layers)
    

jcreinhold

norm = 'batch'
layers = ([res3d_block(1,15,7,norm=norm,dense=True)] +
          [res3d_block(16,16,norm=norm) for _ in range(4)] +
          [conv3d(16,1,ks=1,pad=0,norm=None)])
model = nn.Sequential(*layers)

jcreinhold

loss = nn.MSELoss()
learner = faiv.Learner(idb, model, loss_func=loss)

learner.lr_find()
learner.recorder.plot()  # pick learning rate based on this plot

cbs = [faiv.callbacks.CSVLogger(learner, 'history')]
learner.fit_one_cycle(100, 1e-2, callbacks=cbs)

jcreinhold

jcreinhold

#/usr/bin/env python
# convert all parrec files in a directory to nifti

import argparse
from glob import glob
import os
import sys

import nibabel as nib

jcreinhold

#!/usr/bin/env python
# -*- coding: utf-8 -*-
"""
synthnn.util.metrics

custom metrics for the synthnn package (used in fastai implementation)

Author: Jacob Reinhold (jacob.reinhold@jhu.edu)

Created on: Nov 15, 2018