# ============================================================================================= #
# Author: Lucia Hradecká #
# Copyright: Lucia Hradecká : lucia.d.hradecka@gmail.com #
# #
# MIT License. #
# #
# Permission is hereby granted, free of charge, to any person obtaining a copy #
# of this software and associated documentation files (the "Software"), to deal #
# in the Software without restriction, including without limitation the rights #
# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell #
# copies of the Software, and to permit persons to whom the Software is #
# furnished to do so, subject to the following conditions: #
# #
# The above copyright notice and this permission notice shall be included in all #
# copies or substantial portions of the Software. #
# #
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR #
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, #
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE #
# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER #
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, #
# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE #
# SOFTWARE. #
# ============================================================================================= #
import time
import numpy as np
# bio-volumentations
import bio_volumentations.core.composition as biovol_compose
import bio_volumentations.augmentations as biovol
# TorchIO: requires pytorch, torchIO (pip) (had networkx 2.8.8)
import torchio
# volumentations: require ???
# TODO
# gunpowder: requires gunpowder (pip) (installed networkx 3.2.1)
import gunpowder
# torchvision: requires pytorch, torchvision
import torchvision.transforms.v2 as torchvision_v2
# albumentations: require albumentations
import albumentations
libs = ['biovol', 'torchio', 'volum', 'gunpowder', 'torchvision', 'album']
# image_shape_list = [(1, 256, 256, 256), (3, 256, 256, 256), (1, 256, 256, 256, 10), (3, 256, 256, 256, 10)]
image_shape_list = [(1, 128, 128, 128)]
# num_repeat = 100
num_repeat = 10
augmentations_biovol = [
biovol.RandomCrop(),
biovol.Pad(),
biovol.RandomFlip(axes_to_choose=None, p=1),
biovol.RandomAffineTransform(angle_limit=[22.5, 22.5, 22.5], translation_limit=[10, 10, 10],
scaling_limit=[.2, .2, .2], spacing=[1, 0.5, 2], p=1),
biovol.RandomGaussianBlur(max_sigma=0.8, p=1),
biovol.GaussianNoise(var_limit=(0.001, 0.1), mean=0, p=1),
biovol.RandomBrightnessContrast(brightness_limit=0.2, contrast_limit=0.2, p=1),
biovol.Normalize(mean=0, std=1, p=1),
]
augmentations_torchio = [
torchio.Crop(p=1), # center or random? maybe CroporPad?
torchio.Pad(p=1), # maybe CroporPad?
torchio.CropOrPad(p=1),
torchio.RandomFlip(axes=[], flip_probability=1.0, p=1),
torchio.RandomAffine(scales=[], degrees=[], translation=[], p=1), # TODO isotropic
torchio.RandomBlur(p=1),
torchio.RandomNoise(p=1),
# MISSING: brightness contrast
torchio.ZNormalization(p=1)
]
def get_transformation_list(lib):
if lib == 'biovol':
return augmentations_biovol
if lib == 'torchio':
return augmentations_torchio
if lib == 'volum':
pass
if lib == 'gunpowder':
pass
if lib == 'torchvision':
pass
if lib == 'album':
pass
def init_compose(lib, transform):
if lib == 'biovol':
return biovol_compose.Compose(transforms=[transform], p=1)
if lib == 'torchio':
return torchio.Compose([transform], p=1)
if lib == 'volum':
pass
if lib == 'gunpowder':
pass
if lib == 'torchvision':
pass
if lib == 'album':
pass
def get_input_data(lib, shape):
# Shape is given as ([C], D, H, W, [T])
if lib == 'biovol':
# ([C], D, H, W, [T])
return {'image': np.random.uniform(low=0, high=1, size=shape)}
if lib == 'torchio':
# (C, W, H, D)
return np.random.uniform(low=0, high=1, size=shape.transpose((0, 3, 2, 1)))
if lib == 'volum':
pass
if lib == 'gunpowder':
pass
if lib == 'torchvision':
pass
if lib == 'album':
pass
def transform_data(lib, data, pipeline):
if lib == 'biovol':
augm_data = pipeline(**data)
return augm_data['image'].shape # do something to enforce performing the action
if lib == 'torchio':
pass
if lib == 'volum':
pass
if lib == 'gunpowder':
pass
if lib == 'torchvision':
pass
if lib == 'album':
pass
def single_transform(iterations, size, augmentation, lib):
cumulative = 0
maximum = 0
for i in range(iterations):
test = np.random.uniform(low=0, high=1, size=size)
transformation_pipeline = init_compose(lib, augmentation)
data = {'image': test}
t_0 = time.time()
_ = transform_data(lib, data, transformation_pipeline)
time_spent = time.time() - t_0
cumulative += time_spent
if time_spent > maximum:
maximum = time_spent
return maximum, cumulative
def transformation_speed_benchmark(iterations):
f = open(f"./runtime-{num_repeat}_iterations-all.txt", "w")
for lib in libs:
print(f'*************** LIBRARY {lib} ***************')
for i, augmentation in enumerate(get_transformation_list(lib)):
aug_name = augmentation.__class__.__name__
print(aug_name)
for shape in image_shape_list:
# prepare data and transformation pipeline
transformation_pipeline = init_compose(lib, augmentation)
data = get_input_data(lib, shape)
# run once (prepare the environment)
first_time = time.time()
_ = transform_data(lib, data, transformation_pipeline)
first_result = time.time() - first_time
# run the measured tries
maximum, cumulative = single_transform(iterations, shape, augmentation, lib)
result_time = cumulative / iterations
# log
log_message = f"Runtime in seconds. " \
f"FirstRun: {first_result:.3f}, Average: {result_time:.3f}, Maximum: {maximum:.3f}. " \
f"(Library: {lib}, Transform: {aug_name}, Iterations: {iterations}, ImageSize: {shape})\n"
f.write(log_message)
print(log_message)
print()
f.close()
if __name__ == '__main__':
transformation_speed_benchmark(num_repeat)