t1qc.py

import nibabel as nib
import argparse as ap

import numpy as np

from keras.models import load_model

from dltk.core.io.preprocessing import normalise_zero_one, resize_image_with_crop_or_pad

import keras.backend as K
# from vis.utils import utils

if __name__ == "__main__":

    parser = ap.ArgumentParser(description="Tests an MRI image for motion and other artifacts and returns a probability of success.")

    parser.add_argument("t1image")
    args, leftovers = parser.parse_known_args()

    image_file = args.t1image
    print('Input image file:', image_file)


    # K.set_image_dim_ordering('th')

    try:
        img = nib.load(image_file).get_data()

        img = resize_image_with_crop_or_pad(img, (168, 256, 244), mode='constant')
        img = normalise_zero_one(img)

        # (x_size, y_size, z_size) = img.shape
        # # print('original image size:', img.shape)
        #
        # y_max, z_max = 256, 224
        #
        # while y_size < y_max or z_size < z_max:
        #     img = np.pad(img, 1, 'constant')
        #     (x_size, y_size, z_size) = img.shape
        #
        # y_mid, z_mid = y_size/2, z_size/2
        #
        # x_slice = int(x_size/2)
        # y_start, y_stop = int(y_mid-y_max/2), int(y_mid+y_max/2)
        # z_start, z_stop = int(z_mid-z_max/2), int(z_mid+z_max/2)
        #
        # print('x_slice:', x_slice)
        # print('y_start, y_stop:', y_start, y_stop)
        # print('z_start, z_stop:', z_start, z_stop)
        #
        #
        # img_slice = img[x_slice, y_start:y_stop, z_start:z_stop][np.newaxis, np.newaxis, ...]
        #
        # print(img_slice.shape)

        model = load_model('~/ibis-qc.hdf5')

        slice_predictions = []
        for i in range(10):
            slice_predictions.append(model.predict(img[80+i, :, :][np.newaxis, ..., np.newaxis])[0][0])

        print(np.mean(slice_predictions))

        K.clear_session()

    except Exception as e:
        print(e)