u_volume.append(u_i * norm)
target_volume.append(feed_dict[data.target][0] * norm)
zf_volume.append(feed_dict[data.u][0] * norm)
# postprocessing
u_volume = mriutils.postprocess(np.asarray(u_volume),
data_config['dataset']['name'])
target_volume = mriutils.postprocess(np.asarray(target_volume),
data_config['dataset']['name'])
zf_volume = mriutils.postprocess(np.asarray(zf_volume),
data_config['dataset']['name'])
# evaluation
rmse_vn = mriutils.rmse(u_volume, target_volume)
rmse_zf = mriutils.rmse(zf_volume, target_volume)
ssim_vn = mriutils.ssim(u_volume, target_volume)
ssim_zf = mriutils.ssim(zf_volume, target_volume)
print(
"Zero filling: RMSE={:.4f} SSIM={:.4f} VN: RMSE={:.4f} SSIM={:.4f}"
.format(rmse_zf, ssim_zf, rmse_vn, ssim_vn))
time_reco = time.time() - eval_start_time
print('reconstruction of {} image took {:f}s'.format(
u_volume.shape, time_reco))
print('saving reconstructed image to "{}"'.format(output_name))
# save mat files
patient_id = '%s-p%d' % (data_config['dataset']['name'],
data_config['dataset']['patient'])
mriutils.saveAsMat(u_volume,
# re-normalize images
output.append(u_i[0] * norm)
target.append(feed_dict[data.target][0] * norm)
input0.append(feed_dict[data.u][0] * norm)
normalization.append(norm)
# postprocess images
output = mriutils.postprocess(np.asarray(output),
dataset['name'])
target = mriutils.postprocess(np.asarray(target),
dataset['name'])
input0 = mriutils.postprocess(np.asarray(input0),
dataset['name'])
# evaluation
ssim_patient = mriutils.ssim(output, target)
rmse_patient = mriutils.rmse(output, target)
ssim_eval_dataset.append(ssim_patient)
rmse_eval_dataset.append(rmse_patient)
print(" Patient {:d}: {:8.4f} {:8.4f}".format(
patient, rmse_patient, ssim_patient))
output_path = '%s/%s/%d/' % (eval_output_dir,
dataset['name'], patient)
mriutils.saveAsMat(
output,
'%s/vn-%d.mat' % (output_path, epoch),
'result_vn',
mat_dict={'normalization': np.asarray(normalization)})
mriutils.saveAsMat(
u_i = sess.run(u_op, feed_dict=feed_dict)[0]
u_i = u_i * norm
u_i = mriutils.postprocess(u_i, data_config['dataset']['name'])
# target
target = feed_dict[data.target][0]*norm
target = mriutils.postprocess(target, data_config['dataset']['name'])
# zero filling
zero_filling = feed_dict[data.u][0]*norm
zero_filling = mriutils.postprocess(zero_filling, data_config['dataset']['name'])
# evaluation
rmse_vn = mriutils.rmse(u_i, target)
rmse_zf = mriutils.rmse(zero_filling, target)
ssim_vn = mriutils.ssim(u_i, target)
ssim_zf = mriutils.ssim(zero_filling, target)
print("Zero filling: RMSE={:.4f} SSIM={:.4f} VN: RMSE={:.4f} SSIM={:.4f}".format(rmse_zf, ssim_zf, rmse_vn, ssim_vn))
time_reco = time.time() - eval_start_time
print('reconstruction of {} image took {:f}s'.format(u_i.shape, time_reco))
print('saving reconstructed image to "{}"'.format(output_name))
# save mat files
patient_id = '%s-p%d-sl%d' % (data_config['dataset']['name'],
data_config['dataset']['patient'],
data_config['dataset']['slice'])
mriutils.saveAsMat(u_i, '%s-vn-%s' % (output_name, patient_id), 'result_vn',
mat_dict={'normalization': np.asarray(norm)})
mriutils.saveAsMat(zero_filling, '%s-zf-%s' % (output_name, patient_id), 'result_zf',