# coding=utf-8 # Copyright 2020 The TensorFlow Datasets Authors. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. # Lint as: python3 """DAS beamformed phantom images and paired clinical post-processed images.""" from __future__ import absolute_import from __future__ import division from __future__ import print_function import csv import os import numpy as np import tensorflow.compat.v2 as tf import tensorflow_datasets.public_api as tfds _CITATION = """\ @article{DBLP:journals/corr/abs-1908-05782, author = {Ouwen Huang and Will Long and Nick Bottenus and Gregg E. Trahey and Sina Farsiu and Mark L. Palmeri}, title = {MimickNet, Matching Clinical Post-Processing Under Realistic Black-Box Constraints}, journal = {CoRR}, volume = {abs/1908.05782}, year = {2019}, url = {http://arxiv.org/abs/1908.05782}, archivePrefix = {arXiv}, eprint = {1908.05782}, timestamp = {Mon, 19 Aug 2019 13:21:03 +0200}, biburl = {https://dblp.org/rec/bib/journals/corr/abs-1908-05782}, bibsource = {dblp computer science bibliography, https://dblp.org} }""" _DESCRIPTION = """\ DukeUltrasound is an ultrasound dataset collected at Duke University with a Verasonics c52v probe. It contains delay-and-sum (DAS) beamformed data as well as data post-processed with Siemens Dynamic TCE for speckle reduction, contrast enhancement and improvement in conspicuity of anatomical structures. These data were collected with support from the National Institute of Biomedical Imaging and Bioengineering under Grant R01-EB026574 and National Institutes of Health under Grant 5T32GM007171-44. A usage example is avalible [here](https://colab.research.google.com/drive/1R_ARqpWoiHcUQWg1Fxwyx-ZkLi0IZ5qs).""" _DATA_URL = { 'phantom_data': 'https://research.repository.duke.edu/downloads/vt150j912', 'mark_data': 'https://research.repository.duke.edu/downloads/4x51hj56d' } _DEFAULT_SPLITS = { 'train': 'https://research.repository.duke.edu/downloads/tt44pn391', 'test': 'https://research.repository.duke.edu/downloads/zg64tm441', 'validation': 'https://research.repository.duke.edu/downloads/dj52w535x', 'MARK': 'https://research.repository.duke.edu/downloads/wd375w77v', 'A': 'https://research.repository.duke.edu/downloads/nc580n18d', 'B': 'https://research.repository.duke.edu/downloads/7h149q56p' } class DukeUltrasound(tfds.core.GeneratorBasedBuilder): """DAS beamformed phantom images and paired post-processed images.""" VERSION = tfds.core.Version('1.0.0') def __init__(self, custom_csv_splits=None, **kwargs): """custom_csv_splits is a dictionary of { 'name': 'csvpaths'}.""" super(DukeUltrasound, self).__init__(**kwargs) self.custom_csv_splits = custom_csv_splits if custom_csv_splits else {} def _info(self): return tfds.core.DatasetInfo( builder=self, description=_DESCRIPTION, features=tfds.features.FeaturesDict({ 'das': { 'dB': tfds.features.Tensor(shape=(None,), dtype=tf.float32), 'real': tfds.features.Tensor(shape=(None,), dtype=tf.float32), 'imag': tfds.features.Tensor(shape=(None,), dtype=tf.float32) }, 'dtce': tfds.features.Tensor(shape=(None,), dtype=tf.float32), 'f0_hz': tfds.features.Tensor(shape=(), dtype=tf.float32), 'voltage': tfds.features.Tensor(shape=(), dtype=tf.float32), 'focus_cm': tfds.features.Tensor(shape=(), dtype=tf.float32), 'height': tfds.features.Tensor(shape=(), dtype=tf.uint32), 'width': tfds.features.Tensor(shape=(), dtype=tf.uint32), 'initial_radius': tfds.features.Tensor(shape=(), dtype=tf.float32), 'final_radius': tfds.features.Tensor(shape=(), dtype=tf.float32), 'initial_angle': tfds.features.Tensor(shape=(), dtype=tf.float32), 'final_angle': tfds.features.Tensor(shape=(), dtype=tf.float32), 'probe': tfds.features.Tensor(shape=(), dtype=tf.string), 'scanner': tfds.features.Tensor(shape=(), dtype=tf.string), 'target': tfds.features.Tensor(shape=(), dtype=tf.string), 'timestamp_id': tfds.features.Tensor(shape=(), dtype=tf.uint32), 'harmonic': tfds.features.Tensor(shape=(), dtype=tf.bool) }), supervised_keys=('das/dB', 'dtce'), homepage='https://github.com/ouwen/mimicknet', citation=_CITATION) def _split_generators(self, dl_manager): downloads = _DEFAULT_SPLITS.copy() downloads.update(_DATA_URL) dl_paths = dl_manager.download_and_extract(downloads) splits = [ tfds.core.SplitGenerator( # pylint:disable=g-complex-comprehension name=name, gen_kwargs={ 'datapath': { 'mark_data': dl_paths['mark_data'], 'phantom_data': dl_paths['phantom_data'] }, 'csvpath': dl_paths[name] }) for name, _ in _DEFAULT_SPLITS.items() ] for name, csv_path in self.custom_csv_splits.items(): splits.append( tfds.core.SplitGenerator( name=name, gen_kwargs={ 'datapath': dl_paths['data'], 'csvpath': csv_path })) return splits def _generate_examples(self, datapath, csvpath): with tf.io.gfile.GFile(csvpath) as f: reader = csv.DictReader(f) for row in reader: data_key = 'mark_data' if row['target'] == 'mark' else 'phantom_data' filepath = os.path.join(datapath[data_key], row['filename']) matfile = tfds.core.lazy_imports.scipy.io.loadmat( tf.io.gfile.GFile(filepath, 'rb')) iq = np.abs(np.reshape(matfile['iq'], -1)) iq = iq / iq.max() iq = 20 * np.log10(iq) yield row['filename'], { 'das': { 'dB': iq.astype(np.float32), 'real': np.reshape(matfile['iq'], -1).real.astype(np.float32), 'imag': np.reshape(matfile['iq'], -1).imag.astype(np.float32) }, 'dtce': np.reshape(matfile['dtce'], -1).astype(np.float32), 'f0_hz': row['f0'], 'voltage': row['v'], 'focus_cm': row['focus_cm'], 'height': row['axial_samples'], 'width': row['lateral_samples'], 'initial_radius': row['initial_radius'], 'final_radius': row['final_radius'], 'initial_angle': row['initial_angle'], 'final_angle': row['final_angle'], 'probe': row['probe'], 'scanner': row['scanner'], 'target': row['target'], 'timestamp_id': row['timestamp_id'], 'harmonic': row['harm'] }