pmod prerakmody

## get_mask_boundary.py
"""
This method uses the max pooling operation to get the boundary around a binary mask
Can be used in segmentation tasks where one wants to specific a loss for the boundary pixels/voxels
"""

import os
os.environ["TF_CPP_MIN_LOG_LEVEL"] = "3"
import nrrd
import numpy as np
import tensorflow as tf

## kl_tflow.py
import os
os.environ["TF_CPP_MIN_LOG_LEVEL"] = "3"

import tensorflow as tf # v2.4 with CUDA11.0.0 & CuDNN8.0.0
import tensorflow_probability as tfp # v0.12.1

if len(tf.config.list_physical_devices('GPU')):tf.config.experimental.set_memory_growth(tf.config.list_physical_devices('GPU')[0], True)

x = tf.random.normal((1,100,100,100,10)) # B, H,W,D,C
layer = tfp.layers.Convolution3DFlipout(filters=16, kernel_size=1)

## ece.py
"""
Expected Calibration Error for semantic segmentation tasks
"""
import pdb
import nrrd  # pip install pynrrd
import traceback
import numpy as np
from pathlib import Path
import matplotlib.pyplot as plt

## dice_vs_ce_vs_brier.py
import numpy as np
import matplotlib.pyplot as plt

f,axarr = plt.subplots(1,2)

# Step 1
x       = [0.01, 0.02, 0.03, 0.04] + np.arange(0.05,1.05,0.05).tolist()
y_ce    = [-np.log(each) for each in x]
y_focal1 = [-1*(1 - each)**1*np.log(each) for each in x]
y_focal2 = [-1*(1 - each)**2*np.log(each) for each in x]

## conversion.py
import itk
import SimpleITK as sitk

def convert_sitk_to_itk(sitk_img):

    img_array = sitk.GetArrayFromImage(sitk_img)
    itk_img = itk.GetImageFromArray(img_array)
    itk_img.SetOrigin(tuple(sitk_img.GetOrigin()))
    itk_img.SetSpacing(tuple(sitk_img.GetSpacing()))
    itk_img.SetDirection(itk.GetMatrixFromArray(np.reshape(np.array(sitk_img.GetDirection()), [sitk_img.GetDimension()]*2)))

## autoseg_han.csv

          
            Feature
             AnatomyNet
             FocusNet
             StratifiedNet
             DeepMindNet

            
              Input
               3D Volume
               3D volume
               3D Sub-Volume
               3D Sub-Volume

            
              Crop Info
               178×302×225
               N/A
               128x128x64
               512x512x21

            
              Training Data
               38(MICCAI) + 46(Cetuximab) + 177 (Quebec)
               38(MICCAI)
               38(MICCAI)
               663 (Private dataset)

            
              Augmentation
               Affine + Elastic
               N/A
               Scaling
               2D Augmentations

            
              Architecture
               3D UNet-variant
               3D UNet-variant
               P-HNN
               3D UNet-variant

            
              Additional Blocks
               Squeeze-Excitation Residual Blocks
               Squeeze-Excitation Residual Blocks + DenseASPP
               NAS
               Deptwise-Separable Convs

            
              Cascaded Architecture
               No
               Yes (for smaller optic organs)
               Yes (from anchor to small organs)
               No

            
              Downsampling
               Dim/2
               Dim/2
               N/A
               Dim/64

            
              Final Activation
               softmax
               N/A
               N/A
               Sigmoid

## dataset_loop.py
class HaNMICCAI2015Dataset:
    def __init__():
        # initialize map, filter parameters

    def generator():
        # define generator, map and filter functions

batchsize = None       # user defined value
prefetch_buffer = None # user defined value
shuffle_buffer = None  # user defined value

## gist:2104d0bb5f0412361afee1a3902db5a1
def test(num1, num2):
  return num1/num2

num = 1
den = 2

import timeit
t = timeit.timeit(lambda: test(num, den), number=5)

## convert_elastix_to_itk_transform.py
from pathlib import Path

import itk
import SimpleITK as sitk

config = {
    'transform' : {
        'affine': 'AffineTransform'
        , 'bspline': 'BSplineTransform'
    }

## concurrent.py
import time
import tqdm
import numpy as np
import concurrent
import concurrent.futures


def rand_func(param):
    time.sleep(3)
    print (param)
	"""
	This method uses the max pooling operation to get the boundary around a binary mask
	Can be used in segmentation tasks where one wants to specific a loss for the boundary pixels/voxels
	"""

	import os
	os.environ["TF_CPP_MIN_LOG_LEVEL"] = "3"
	import nrrd
	import numpy as np
	import tensorflow as tf
	import os
	os.environ["TF_CPP_MIN_LOG_LEVEL"] = "3"

	import tensorflow as tf # v2.4 with CUDA11.0.0 & CuDNN8.0.0
	import tensorflow_probability as tfp # v0.12.1

	if len(tf.config.list_physical_devices('GPU')):tf.config.experimental.set_memory_growth(tf.config.list_physical_devices('GPU')[0], True)

	x = tf.random.normal((1,100,100,100,10)) # B, H,W,D,C
	layer = tfp.layers.Convolution3DFlipout(filters=16, kernel_size=1)
	"""
	Expected Calibration Error for semantic segmentation tasks
	"""
	import pdb
	import nrrd # pip install pynrrd
	import traceback
	import numpy as np
	from pathlib import Path
	import matplotlib.pyplot as plt
	import numpy as np
	import matplotlib.pyplot as plt

	f,axarr = plt.subplots(1,2)

	# Step 1
	x = [0.01, 0.02, 0.03, 0.04] + np.arange(0.05,1.05,0.05).tolist()
	y_ce = [-np.log(each) for each in x]
	y_focal1 = [-1(1 - each)1np.log(each) for each in x]
	y_focal2 = [-1(1 - each)2np.log(each) for each in x]
	import itk
	import SimpleITK as sitk

	def convert_sitk_to_itk(sitk_img):

	img_array = sitk.GetArrayFromImage(sitk_img)
	itk_img = itk.GetImageFromArray(img_array)
	itk_img.SetOrigin(tuple(sitk_img.GetOrigin()))
	itk_img.SetSpacing(tuple(sitk_img.GetSpacing()))
	itk_img.SetDirection(itk.GetMatrixFromArray(np.reshape(np.array(sitk_img.GetDirection()), [sitk_img.GetDimension()]*2)))
Feature	AnatomyNet	FocusNet	StratifiedNet	DeepMindNet
Input	3D Volume	3D volume	3D Sub-Volume	3D Sub-Volume
Crop Info	178×302×225	N/A	128x128x64	512x512x21
Training Data	38(MICCAI) + 46(Cetuximab) + 177 (Quebec)	38(MICCAI)	38(MICCAI)	663 (Private dataset)
Augmentation	Affine + Elastic	N/A	Scaling	2D Augmentations
Architecture	3D UNet-variant	3D UNet-variant	P-HNN	3D UNet-variant
Additional Blocks	Squeeze-Excitation Residual Blocks	Squeeze-Excitation Residual Blocks + DenseASPP	NAS	Deptwise-Separable Convs
Cascaded Architecture	No	Yes (for smaller optic organs)	Yes (from anchor to small organs)	No
Downsampling	Dim/2	Dim/2	N/A	Dim/64
Final Activation	softmax	N/A	N/A	Sigmoid
	class HaNMICCAI2015Dataset:
	def __init__():
	# initialize map, filter parameters

	def generator():
	# define generator, map and filter functions

	batchsize = None # user defined value
	prefetch_buffer = None # user defined value
	shuffle_buffer = None # user defined value
	def test(num1, num2):
	return num1/num2

	num = 1
	den = 2

	import timeit
	t = timeit.timeit(lambda: test(num, den), number=5)
	import time
	import tqdm
	import numpy as np
	import concurrent
	import concurrent.futures


	def rand_func(param):
	time.sleep(3)
	print (param)