DenseNet and Deconvolution.
A Sparse-View CT Reconstruction Method Based on Combination of DenseNet and Deconvolution
Didn’t get good results.
I am trying to reimplement the paper below in keras
paper: https://ieeexplore.ieee.org/document/8331861 github: https://github.com/zzc623/DD_Net
from google.colab import drive
drive.mount('/content/drive')
Go to this URL in a browser: https://accounts.google.com/o/oauth2/auth?client_id=947318989803-6bn6qk8qdgf4n4g3pfee6491hc0brc4i.apps.googleusercontent.com&redirect_uri=urn%3aietf%3awg%3aoauth%3a2.0%3aoob&response_type=code&scope=email%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdocs.test%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive.photos.readonly%20https%3a%2f%2fwww.googleapis.com%2fauth%2fpeopleapi.readonly
Enter your authorization code:
··········
Mounted at /content/drive
paper: https://ieeexplore.ieee.org/document/8331861 github: https://github.com/zzc623/DD_Net
import os
import keras
import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt
import matplotlib.image as mpimg
from keras.layers import Conv2D, MaxPooling2D, Input, Dense, UpSampling2D
from keras.layers import BatchNormalization, Activation, Dropout, Subtract
from keras.models import Model
from glob import glob
from keras.layers.convolutional import Conv2DTranspose
from keras.layers import concatenate
Using TensorFlow backend.
The default version of TensorFlow in Colab will soon switch to TensorFlow 2.x.
We recommend you upgrade now
or ensure your notebook will continue to use TensorFlow 1.x via the %tensorflow_version 1.x magic:
more info.
# %tensorflow_version 1.x
# import tensorflow as tf
print(tf.__version__)
1.15.0
If we have weights we canload them: loaded_model.load_weights(“model.h5”)
# print(len(os.listdir('/content/drive/My Drive/Colab Notebooks/CT_data/sparseview_60/train')))
# print(len(os.listdir('/content/drive/My Drive/Colab Notebooks/CT_data/ndct/train')))
# # 3600
# # 3600
# print(len(os.listdir('/content/drive/My Drive/Colab Notebooks/CT_data/sparseview_60/test/')))
# print(len(os.listdir('/content/drive/My Drive/Colab Notebooks/CT_data/ndct/test/')))
# # # 354
# # # 354
ndct = sorted(glob('/content/drive/My Drive/Colab Notebooks/CT_data/ndct/train/*'))
ldct = sorted(glob('/content/drive/My Drive/Colab Notebooks/CT_data/sparseview_60/train/*'))
ndct_test = sorted(glob('/content/drive/My Drive/Colab Notebooks/CT_data/ndct/test/*'))
ldct_test = sorted(glob('/content/drive/My Drive/Colab Notebooks/CT_data/sparseview_60/test/*'))
print(len(ndct))
print(len(ldct))
print(len(ndct_test))
print(len(ldct_test))
3600
3600
354
354
The formulas below will be used to calculate the quality of the reconstruction. Higher PSNR generally indicates high quality of reconstruction.
def cal_psnr(im1, im2):
# assert pixel value range is 0-255 and type is uint8
mse = ((im1.astype(np.float) - im2.astype(np.float)) ** 2).mean()
maxval = np.amax(im1)
psnr = 10 * np.log10(maxval ** 2 / mse)
return psnr
def tf_psnr(im1, im2):
# assert pixel value range is 0-1
#mse = tf.losses.mean_squared_error(labels=im2 * 255.0, predictions=im1 * 255.0)
mse = tf.compat.v1.losses.mean_squared_error(labels=im2 * 255.0, predictions=im1 * 255.0)
return 10.0 * (tf.log(255.0 ** 2 / mse) / tf.log(10.0))
Using less data: #for i in range(0, 600). Have 3600 in all. Processing 3600 images takes aprox. 20 minutes to run. But once we create .npy aray we don’t have to rerun this code in the future and we will have .npy form of our data. Colab has 11GB RAM limit.
ndct_imgs_train = []
# for i in range(0, len(ndct)):
for i in range(0, 600):
f = open(ndct[i],'rb')
a = np.fromfile(f, np.float32)
ndct_imgs_train.append(a)
f.close()
print("len(ndct_imgs_train)....: ",len(ndct_imgs_train))
#len(ndct_imgs_train)....: 3600
Using different range to use less data to train #for i in range(0, 600). In all have 3600 images. It takes aprox. 20 minutes to process all 3600. But once we create .npy aray we don’t have to rerun this code in the future and we will have .npy form of our data.
ldct_imgs_train = []
# for i in range(0, len(ldct)):
for i in range(0, 600):
f = open(ldct[i],'rb')
a = np.fromfile(f, np.float32)
ldct_imgs_train.append(a)
f.close()
print("len(ldct_imgs_train)....: ",len(ldct_imgs_train))
only using 100 images to test
ndct_imgs_test = []
# for i in range(0, len(ndct_test)):
for i in range(0, 100):
f = open(ndct_test[i],'rb')
a = np.fromfile(f, np.float32)
ndct_imgs_test.append(a)
f.close()
print("len(ndct_imgs_test)....: ",len(ndct_imgs_test))
only using 100 images to test
# load the image
ldct_imgs_test = []
# for i in range(0, len(ldct_test)):
for i in range(0, 100):
f = open(ldct_test[i],'rb')
a = np.fromfile(f, np.float32)
ldct_imgs_test.append(a)
f.close()
print("len(ldct_imgs_test)....: ",len(ldct_imgs_test))
Must reshape images to train
ldct_train = np.asarray(ldct_imgs_train)
ndct_train = np.asarray(ndct_imgs_train)
ldct_train = ldct_train.reshape(600,512,512,1)
ndct_train = ndct_train.reshape(600,512,512,1)
ldct_test = np.asarray(ldct_imgs_test)
ndct_test = np.asarray(ndct_imgs_test)
# ldct_test = ldct_test.reshape(len(ldct_imgs_test),512,512,1)
# ndct_test = ndct_test.reshape(len(ldct_imgs_test),512,512,1)
ldct_test = ldct_test.reshape(100,512,512,1)
ndct_test = ndct_test.reshape(100,512,512,1)
print(ldct_train.shape)
print(ndct_train.shape)
print(ldct_test.shape)
print(ndct_test.shape)
# np.save('sparseview_60_train_600', ldct_train) # save the file as "sparseview_60_train.npy"
# np.save('ndct_train_600', ndct_train) # save the file as "ndct_train.npy"
# np.save('sparseview_60_test_100', ldct_test) # save the file as "sparseview_60_test.npy"
# np.save('ndct_test_100', ndct_test) # save the file as "ndct_test.npy"
np.save('/content/drive/My Drive/Colab Notebooks/dd_net/sparseview_60_train_600', ldct_train) # save the file as "sparseview_60_train.npy"
np.save('/content/drive/My Drive/Colab Notebooks/dd_net/ndct_train_600', ndct_train) # save the file as "ndct_train.npy"
np.save('/content/drive/My Drive/Colab Notebooks/dd_net/sparseview_60_test_100', ldct_test) # save the file as "sparseview_60_test.npy"
np.save('/content/drive/My Drive/Colab Notebooks/dd_net/ndct_test_100', ndct_test) # save the file as "ndct_test.npy"
sparseview_60_train = np.load('/content/drive/My Drive/Colab Notebooks/dd_net/sparseview_60_train_600.npy') # loads saved array into variable sparseview_60_train.
ndct_train = np.load('/content/drive/My Drive/Colab Notebooks/dd_net/ndct_train_600.npy') # loads saved array into variable ndct_train.
sparseview_60_test = np.load('/content/drive/My Drive/Colab Notebooks/dd_net/sparseview_60_test_100.npy') # loads saved array into variable sparseview_60_test.
ndct_test = np.load('/content/drive/My Drive/Colab Notebooks/dd_net/ndct_test_100.npy') # loads saved array into variable ndct_test.
# sparseview_60_train = np.load('sparseview_60_train_600.npy') # loads saved array into variable sparseview_60_train.
# ndct_train = np.load('ndct_train_600.npy') # loads saved array into variable ndct_train.
# sparseview_60_test = np.load('sparseview_60_test_100.npy') # loads saved array into variable sparseview_60_test.
# ndct_test = np.load('ndct_test_100.npy') # loads saved array into variable ndct_test.
def denseblock(input):
# - L1
d2_1 = BatchNormalization()(input)
d2_1 = Activation('relu')(d2_1)
d2_1 = Conv2D(16, (5, 5), padding='same', use_bias=False, strides=(1, 1))(d2_1)
d2_1 = concatenate([input, d2_1])
# - L2
d2_2 = BatchNormalization()(d2_1)
d2_2 = Activation('relu')(d2_2)
d2_2= Conv2D(16, (5, 5), padding='same', use_bias=False, strides=(1, 1))(d2_2)
d2_2 = concatenate([input, d2_1, d2_2])
# - L3
d2_3 = BatchNormalization()(d2_2)
d2_3 = Activation('relu')(d2_3)
d2_3 = Conv2D(16, (5, 5), padding='same', use_bias=False, strides=(1, 1))(d2_3)
d2_3 = concatenate([input, d2_1, d2_2, d2_3])
# - L4
d2_4 = Conv2D(16, (5, 5), padding='same', use_bias=False, strides=(1, 1))(d2_3)
d2_4 = BatchNormalization()(d2_4)
d2_4 = Activation('relu')(d2_4)
d2_4= Conv2D(16, (5, 5), padding='same', use_bias=False, strides=(1, 1))(d2_4)
d2_4 = concatenate([input, d2_1, d2_2, d2_3, d2_4])
return d2_4
inputs = Input((None, None,1))
# ---A1 Layer-----------------------
a1 = Conv2D(16, (7, 7), padding='same', use_bias=False, strides=(1, 1))(inputs)
a1 = MaxPooling2D((2, 2)) (a1)
# images 256 X 256
d1 = denseblock(a1)
a1 = BatchNormalization()(d1)
a1 = Activation('relu')(a1)
a1 = Conv2D(16, (1, 1), strides=(1, 1)) (a1)
# ----A2 Layer---------------------
a2 = MaxPooling2D((2, 2)) (a1)
# images 128 X 128
d2 = denseblock(a2)
a2 = BatchNormalization()(d2)
a2 = Activation('relu')(a2)
a2 = Conv2D(16, (1, 1), strides=(1, 1)) (a2)
# # ----A3 Layer----------------------
a3 = MaxPooling2D((2, 2)) (a2)
# images 64 X 64
d3 = denseblock(a3)
a3 = BatchNormalization()(d3)
a3 = Activation('relu')(a3)
a3 = Conv2D(16, (1, 1), strides=(1, 1)) (a3)
# ----A4 Layer----------------------
a4 = MaxPooling2D((2, 2)) (a3)
# images 32 X 3
d4 = denseblock(a4)
a4 = BatchNormalization()(d4)
a4 = Activation('relu')(a4)
a4 = Conv2D(16, (1, 1), strides=(1, 1)) (a4)
# #----B1 Layer-----------------------
b1 = UpSampling2D((2, 2)) (a4)
# images 64 X 64
b1 = concatenate([b1, a3])
b1 = Conv2DTranspose(16, (5, 5), padding='same', strides=(1, 1)) (b1)
b1 = Activation('relu')(b1)
b1 = BatchNormalization()(b1)
b1 = Conv2DTranspose(16, (1, 1), padding='same', strides=(1, 1)) (b1)
b1 = Activation('relu')(b1)
b1 = BatchNormalization()(b1)
# #----B2 Layer-----------------------
b2 = UpSampling2D((2, 2)) (b1)
# images 128 X 128
b2 = concatenate([b2, a2])
b2 = Conv2DTranspose(16, (5, 5), padding='same', strides=(1, 1)) (b2)
b2 = Activation('relu')(b2)
b2 = BatchNormalization()(b2)
b2 = Conv2DTranspose(16, (1, 1), padding='same', strides=(1, 1)) (b2)
b2 = Activation('relu')(b2)
b2 = BatchNormalization()(b2)
#----B3 Layer------------------------
b3 = UpSampling2D((2, 2)) (b2)
# images 256 X 256
b3 = concatenate([b3, a1])
b3 = Conv2DTranspose(16, (5, 5), padding='same', strides=(1, 1)) (b3)
b3 = Activation('relu')(b3)
b3 = BatchNormalization()(b3)
b3 = Conv2DTranspose(16, (1, 1), padding='same', strides=(1, 1)) (b3)
b3 = Activation('relu')(b3)
b3 = BatchNormalization()(b3)
#----B4 Layer-------------------------
b4 = UpSampling2D((2, 2)) (b3)
# images 512 X 512
b4 = concatenate([b4, inputs])
b4 = Conv2DTranspose(16, (5, 5),padding='same', strides=(1, 1)) (b4)
b4 = Activation('relu')(b4)
output_img = Conv2DTranspose(1, (1, 1), strides=(1, 1)) (b4)
# output_img = Activation('relu')(b4)
# ------ end B4 layer
subtracted = Subtract()([inputs, output_img])
ddnet_model = Model(inputs=[inputs], outputs=[subtracted])
ddnet_model.compile(optimizer='adam', loss='mse', metrics=[tf_psnr])
DDNet model
ddnet_model.summary()
Model: "model_2"
__________________________________________________________________________________________________
Layer (type) Output Shape Param # Connected to
==================================================================================================
input_2 (InputLayer) (None, None, None, 1 0
__________________________________________________________________________________________________
conv2d_26 (Conv2D) (None, None, None, 1 784 input_2[0][0]
__________________________________________________________________________________________________
max_pooling2d_5 (MaxPooling2D) (None, None, None, 1 0 conv2d_26[0][0]
__________________________________________________________________________________________________
batch_normalization_27 (BatchNo (None, None, None, 1 64 max_pooling2d_5[0][0]
__________________________________________________________________________________________________
activation_28 (Activation) (None, None, None, 1 0 batch_normalization_27[0][0]
__________________________________________________________________________________________________
conv2d_27 (Conv2D) (None, None, None, 1 6400 activation_28[0][0]
__________________________________________________________________________________________________
concatenate_21 (Concatenate) (None, None, None, 3 0 max_pooling2d_5[0][0]
conv2d_27[0][0]
__________________________________________________________________________________________________
batch_normalization_28 (BatchNo (None, None, None, 3 128 concatenate_21[0][0]
__________________________________________________________________________________________________
activation_29 (Activation) (None, None, None, 3 0 batch_normalization_28[0][0]
__________________________________________________________________________________________________
conv2d_28 (Conv2D) (None, None, None, 1 12800 activation_29[0][0]
__________________________________________________________________________________________________
concatenate_22 (Concatenate) (None, None, None, 6 0 max_pooling2d_5[0][0]
concatenate_21[0][0]
conv2d_28[0][0]
__________________________________________________________________________________________________
batch_normalization_29 (BatchNo (None, None, None, 6 256 concatenate_22[0][0]
__________________________________________________________________________________________________
activation_30 (Activation) (None, None, None, 6 0 batch_normalization_29[0][0]
__________________________________________________________________________________________________
conv2d_29 (Conv2D) (None, None, None, 1 25600 activation_30[0][0]
__________________________________________________________________________________________________
concatenate_23 (Concatenate) (None, None, None, 1 0 max_pooling2d_5[0][0]
concatenate_21[0][0]
concatenate_22[0][0]
conv2d_29[0][0]
__________________________________________________________________________________________________
conv2d_30 (Conv2D) (None, None, None, 1 51200 concatenate_23[0][0]
__________________________________________________________________________________________________
batch_normalization_30 (BatchNo (None, None, None, 1 64 conv2d_30[0][0]
__________________________________________________________________________________________________
activation_31 (Activation) (None, None, None, 1 0 batch_normalization_30[0][0]
__________________________________________________________________________________________________
conv2d_31 (Conv2D) (None, None, None, 1 6400 activation_31[0][0]
__________________________________________________________________________________________________
concatenate_24 (Concatenate) (None, None, None, 2 0 max_pooling2d_5[0][0]
concatenate_21[0][0]
concatenate_22[0][0]
concatenate_23[0][0]
conv2d_31[0][0]
__________________________________________________________________________________________________
batch_normalization_31 (BatchNo (None, None, None, 2 1024 concatenate_24[0][0]
__________________________________________________________________________________________________
activation_32 (Activation) (None, None, None, 2 0 batch_normalization_31[0][0]
__________________________________________________________________________________________________
conv2d_32 (Conv2D) (None, None, None, 1 4112 activation_32[0][0]
__________________________________________________________________________________________________
max_pooling2d_6 (MaxPooling2D) (None, None, None, 1 0 conv2d_32[0][0]
__________________________________________________________________________________________________
batch_normalization_32 (BatchNo (None, None, None, 1 64 max_pooling2d_6[0][0]
__________________________________________________________________________________________________
activation_33 (Activation) (None, None, None, 1 0 batch_normalization_32[0][0]
__________________________________________________________________________________________________
conv2d_33 (Conv2D) (None, None, None, 1 6400 activation_33[0][0]
__________________________________________________________________________________________________
concatenate_25 (Concatenate) (None, None, None, 3 0 max_pooling2d_6[0][0]
conv2d_33[0][0]
__________________________________________________________________________________________________
batch_normalization_33 (BatchNo (None, None, None, 3 128 concatenate_25[0][0]
__________________________________________________________________________________________________
activation_34 (Activation) (None, None, None, 3 0 batch_normalization_33[0][0]
__________________________________________________________________________________________________
conv2d_34 (Conv2D) (None, None, None, 1 12800 activation_34[0][0]
__________________________________________________________________________________________________
concatenate_26 (Concatenate) (None, None, None, 6 0 max_pooling2d_6[0][0]
concatenate_25[0][0]
conv2d_34[0][0]
__________________________________________________________________________________________________
batch_normalization_34 (BatchNo (None, None, None, 6 256 concatenate_26[0][0]
__________________________________________________________________________________________________
activation_35 (Activation) (None, None, None, 6 0 batch_normalization_34[0][0]
__________________________________________________________________________________________________
conv2d_35 (Conv2D) (None, None, None, 1 25600 activation_35[0][0]
__________________________________________________________________________________________________
concatenate_27 (Concatenate) (None, None, None, 1 0 max_pooling2d_6[0][0]
concatenate_25[0][0]
concatenate_26[0][0]
conv2d_35[0][0]
__________________________________________________________________________________________________
conv2d_36 (Conv2D) (None, None, None, 1 51200 concatenate_27[0][0]
__________________________________________________________________________________________________
batch_normalization_35 (BatchNo (None, None, None, 1 64 conv2d_36[0][0]
__________________________________________________________________________________________________
activation_36 (Activation) (None, None, None, 1 0 batch_normalization_35[0][0]
__________________________________________________________________________________________________
conv2d_37 (Conv2D) (None, None, None, 1 6400 activation_36[0][0]
__________________________________________________________________________________________________
concatenate_28 (Concatenate) (None, None, None, 2 0 max_pooling2d_6[0][0]
concatenate_25[0][0]
concatenate_26[0][0]
concatenate_27[0][0]
conv2d_37[0][0]
__________________________________________________________________________________________________
batch_normalization_36 (BatchNo (None, None, None, 2 1024 concatenate_28[0][0]
__________________________________________________________________________________________________
activation_37 (Activation) (None, None, None, 2 0 batch_normalization_36[0][0]
__________________________________________________________________________________________________
conv2d_38 (Conv2D) (None, None, None, 1 4112 activation_37[0][0]
__________________________________________________________________________________________________
max_pooling2d_7 (MaxPooling2D) (None, None, None, 1 0 conv2d_38[0][0]
__________________________________________________________________________________________________
batch_normalization_37 (BatchNo (None, None, None, 1 64 max_pooling2d_7[0][0]
__________________________________________________________________________________________________
activation_38 (Activation) (None, None, None, 1 0 batch_normalization_37[0][0]
__________________________________________________________________________________________________
conv2d_39 (Conv2D) (None, None, None, 1 6400 activation_38[0][0]
__________________________________________________________________________________________________
concatenate_29 (Concatenate) (None, None, None, 3 0 max_pooling2d_7[0][0]
conv2d_39[0][0]
__________________________________________________________________________________________________
batch_normalization_38 (BatchNo (None, None, None, 3 128 concatenate_29[0][0]
__________________________________________________________________________________________________
activation_39 (Activation) (None, None, None, 3 0 batch_normalization_38[0][0]
__________________________________________________________________________________________________
conv2d_40 (Conv2D) (None, None, None, 1 12800 activation_39[0][0]
__________________________________________________________________________________________________
concatenate_30 (Concatenate) (None, None, None, 6 0 max_pooling2d_7[0][0]
concatenate_29[0][0]
conv2d_40[0][0]
__________________________________________________________________________________________________
batch_normalization_39 (BatchNo (None, None, None, 6 256 concatenate_30[0][0]
__________________________________________________________________________________________________
activation_40 (Activation) (None, None, None, 6 0 batch_normalization_39[0][0]
__________________________________________________________________________________________________
conv2d_41 (Conv2D) (None, None, None, 1 25600 activation_40[0][0]
__________________________________________________________________________________________________
concatenate_31 (Concatenate) (None, None, None, 1 0 max_pooling2d_7[0][0]
concatenate_29[0][0]
concatenate_30[0][0]
conv2d_41[0][0]
__________________________________________________________________________________________________
conv2d_42 (Conv2D) (None, None, None, 1 51200 concatenate_31[0][0]
__________________________________________________________________________________________________
batch_normalization_40 (BatchNo (None, None, None, 1 64 conv2d_42[0][0]
__________________________________________________________________________________________________
activation_41 (Activation) (None, None, None, 1 0 batch_normalization_40[0][0]
__________________________________________________________________________________________________
conv2d_43 (Conv2D) (None, None, None, 1 6400 activation_41[0][0]
__________________________________________________________________________________________________
concatenate_32 (Concatenate) (None, None, None, 2 0 max_pooling2d_7[0][0]
concatenate_29[0][0]
concatenate_30[0][0]
concatenate_31[0][0]
conv2d_43[0][0]
__________________________________________________________________________________________________
batch_normalization_41 (BatchNo (None, None, None, 2 1024 concatenate_32[0][0]
__________________________________________________________________________________________________
activation_42 (Activation) (None, None, None, 2 0 batch_normalization_41[0][0]
__________________________________________________________________________________________________
conv2d_44 (Conv2D) (None, None, None, 1 4112 activation_42[0][0]
__________________________________________________________________________________________________
max_pooling2d_8 (MaxPooling2D) (None, None, None, 1 0 conv2d_44[0][0]
__________________________________________________________________________________________________
batch_normalization_42 (BatchNo (None, None, None, 1 64 max_pooling2d_8[0][0]
__________________________________________________________________________________________________
activation_43 (Activation) (None, None, None, 1 0 batch_normalization_42[0][0]
__________________________________________________________________________________________________
conv2d_45 (Conv2D) (None, None, None, 1 6400 activation_43[0][0]
__________________________________________________________________________________________________
concatenate_33 (Concatenate) (None, None, None, 3 0 max_pooling2d_8[0][0]
conv2d_45[0][0]
__________________________________________________________________________________________________
batch_normalization_43 (BatchNo (None, None, None, 3 128 concatenate_33[0][0]
__________________________________________________________________________________________________
activation_44 (Activation) (None, None, None, 3 0 batch_normalization_43[0][0]
__________________________________________________________________________________________________
conv2d_46 (Conv2D) (None, None, None, 1 12800 activation_44[0][0]
__________________________________________________________________________________________________
concatenate_34 (Concatenate) (None, None, None, 6 0 max_pooling2d_8[0][0]
concatenate_33[0][0]
conv2d_46[0][0]
__________________________________________________________________________________________________
batch_normalization_44 (BatchNo (None, None, None, 6 256 concatenate_34[0][0]
__________________________________________________________________________________________________
activation_45 (Activation) (None, None, None, 6 0 batch_normalization_44[0][0]
__________________________________________________________________________________________________
conv2d_47 (Conv2D) (None, None, None, 1 25600 activation_45[0][0]
__________________________________________________________________________________________________
concatenate_35 (Concatenate) (None, None, None, 1 0 max_pooling2d_8[0][0]
concatenate_33[0][0]
concatenate_34[0][0]
conv2d_47[0][0]
__________________________________________________________________________________________________
conv2d_48 (Conv2D) (None, None, None, 1 51200 concatenate_35[0][0]
__________________________________________________________________________________________________
batch_normalization_45 (BatchNo (None, None, None, 1 64 conv2d_48[0][0]
__________________________________________________________________________________________________
activation_46 (Activation) (None, None, None, 1 0 batch_normalization_45[0][0]
__________________________________________________________________________________________________
conv2d_49 (Conv2D) (None, None, None, 1 6400 activation_46[0][0]
__________________________________________________________________________________________________
concatenate_36 (Concatenate) (None, None, None, 2 0 max_pooling2d_8[0][0]
concatenate_33[0][0]
concatenate_34[0][0]
concatenate_35[0][0]
conv2d_49[0][0]
__________________________________________________________________________________________________
batch_normalization_46 (BatchNo (None, None, None, 2 1024 concatenate_36[0][0]
__________________________________________________________________________________________________
activation_47 (Activation) (None, None, None, 2 0 batch_normalization_46[0][0]
__________________________________________________________________________________________________
conv2d_50 (Conv2D) (None, None, None, 1 4112 activation_47[0][0]
__________________________________________________________________________________________________
up_sampling2d_5 (UpSampling2D) (None, None, None, 1 0 conv2d_50[0][0]
__________________________________________________________________________________________________
concatenate_37 (Concatenate) (None, None, None, 3 0 up_sampling2d_5[0][0]
conv2d_44[0][0]
__________________________________________________________________________________________________
conv2d_transpose_9 (Conv2DTrans (None, None, None, 1 12816 concatenate_37[0][0]
__________________________________________________________________________________________________
activation_48 (Activation) (None, None, None, 1 0 conv2d_transpose_9[0][0]
__________________________________________________________________________________________________
batch_normalization_47 (BatchNo (None, None, None, 1 64 activation_48[0][0]
__________________________________________________________________________________________________
conv2d_transpose_10 (Conv2DTran (None, None, None, 1 272 batch_normalization_47[0][0]
__________________________________________________________________________________________________
activation_49 (Activation) (None, None, None, 1 0 conv2d_transpose_10[0][0]
__________________________________________________________________________________________________
batch_normalization_48 (BatchNo (None, None, None, 1 64 activation_49[0][0]
__________________________________________________________________________________________________
up_sampling2d_6 (UpSampling2D) (None, None, None, 1 0 batch_normalization_48[0][0]
__________________________________________________________________________________________________
concatenate_38 (Concatenate) (None, None, None, 3 0 up_sampling2d_6[0][0]
conv2d_38[0][0]
__________________________________________________________________________________________________
conv2d_transpose_11 (Conv2DTran (None, None, None, 1 12816 concatenate_38[0][0]
__________________________________________________________________________________________________
activation_50 (Activation) (None, None, None, 1 0 conv2d_transpose_11[0][0]
__________________________________________________________________________________________________
batch_normalization_49 (BatchNo (None, None, None, 1 64 activation_50[0][0]
__________________________________________________________________________________________________
conv2d_transpose_12 (Conv2DTran (None, None, None, 1 272 batch_normalization_49[0][0]
__________________________________________________________________________________________________
activation_51 (Activation) (None, None, None, 1 0 conv2d_transpose_12[0][0]
__________________________________________________________________________________________________
batch_normalization_50 (BatchNo (None, None, None, 1 64 activation_51[0][0]
__________________________________________________________________________________________________
up_sampling2d_7 (UpSampling2D) (None, None, None, 1 0 batch_normalization_50[0][0]
__________________________________________________________________________________________________
concatenate_39 (Concatenate) (None, None, None, 3 0 up_sampling2d_7[0][0]
conv2d_32[0][0]
__________________________________________________________________________________________________
conv2d_transpose_13 (Conv2DTran (None, None, None, 1 12816 concatenate_39[0][0]
__________________________________________________________________________________________________
activation_52 (Activation) (None, None, None, 1 0 conv2d_transpose_13[0][0]
__________________________________________________________________________________________________
batch_normalization_51 (BatchNo (None, None, None, 1 64 activation_52[0][0]
__________________________________________________________________________________________________
conv2d_transpose_14 (Conv2DTran (None, None, None, 1 272 batch_normalization_51[0][0]
__________________________________________________________________________________________________
activation_53 (Activation) (None, None, None, 1 0 conv2d_transpose_14[0][0]
__________________________________________________________________________________________________
batch_normalization_52 (BatchNo (None, None, None, 1 64 activation_53[0][0]
__________________________________________________________________________________________________
up_sampling2d_8 (UpSampling2D) (None, None, None, 1 0 batch_normalization_52[0][0]
__________________________________________________________________________________________________
concatenate_40 (Concatenate) (None, None, None, 1 0 up_sampling2d_8[0][0]
input_2[0][0]
__________________________________________________________________________________________________
conv2d_transpose_15 (Conv2DTran (None, None, None, 1 6816 concatenate_40[0][0]
__________________________________________________________________________________________________
activation_54 (Activation) (None, None, None, 1 0 conv2d_transpose_15[0][0]
__________________________________________________________________________________________________
conv2d_transpose_16 (Conv2DTran (None, None, None, 1 17 activation_54[0][0]
__________________________________________________________________________________________________
subtract_2 (Subtract) (None, None, None, 1 0 input_2[0][0]
conv2d_transpose_16[0][0]
==================================================================================================
Total params: 479,457
Trainable params: 476,193
Non-trainable params: 3,264
__________________________________________________________________________________________________
history=ddnet_model.fit(sparseview_60_train, ndct_train, validation_split=0.1, batch_size=10, epochs=40)
Train on 540 samples, validate on 60 samples
Epoch 1/40
540/540 [==============================] - 19s 35ms/step - loss: 0.0011 - tf_psnr: 29.8691 - val_loss: 0.0019 - val_tf_psnr: 27.2116
Epoch 2/40
540/540 [==============================] - 19s 35ms/step - loss: 0.0011 - tf_psnr: 29.7986 - val_loss: 0.0017 - val_tf_psnr: 27.6557
Epoch 3/40
540/540 [==============================] - 19s 35ms/step - loss: 7.9521e-04 - tf_psnr: 31.0269 - val_loss: 0.0010 - val_tf_psnr: 29.9420
Epoch 4/40
540/540 [==============================] - 19s 35ms/step - loss: 8.5768e-04 - tf_psnr: 30.7284 - val_loss: 7.3941e-04 - val_tf_psnr: 31.3242
Epoch 5/40
540/540 [==============================] - 19s 35ms/step - loss: 7.7062e-04 - tf_psnr: 31.1725 - val_loss: 7.0704e-04 - val_tf_psnr: 31.5166
Epoch 6/40
540/540 [==============================] - 19s 35ms/step - loss: 8.8858e-04 - tf_psnr: 30.6390 - val_loss: 7.9605e-04 - val_tf_psnr: 30.9996
Epoch 7/40
540/540 [==============================] - 19s 35ms/step - loss: 7.2811e-04 - tf_psnr: 31.4067 - val_loss: 6.8186e-04 - val_tf_psnr: 31.6725
Epoch 8/40
540/540 [==============================] - 19s 35ms/step - loss: 6.9759e-04 - tf_psnr: 31.5844 - val_loss: 8.0168e-04 - val_tf_psnr: 30.9784
Epoch 9/40
540/540 [==============================] - 19s 35ms/step - loss: 8.2150e-04 - tf_psnr: 30.9244 - val_loss: 0.0016 - val_tf_psnr: 28.1129
Epoch 10/40
540/540 [==============================] - 19s 35ms/step - loss: 8.3867e-04 - tf_psnr: 30.8462 - val_loss: 7.0182e-04 - val_tf_psnr: 31.5452
Epoch 11/40
540/540 [==============================] - 19s 35ms/step - loss: 6.6414e-04 - tf_psnr: 31.8029 - val_loss: 6.0881e-04 - val_tf_psnr: 32.1663
Epoch 12/40
540/540 [==============================] - 19s 35ms/step - loss: 6.9797e-04 - tf_psnr: 31.5900 - val_loss: 8.1600e-04 - val_tf_psnr: 30.8983
Epoch 13/40
540/540 [==============================] - 19s 35ms/step - loss: 7.3342e-04 - tf_psnr: 31.3820 - val_loss: 6.0540e-04 - val_tf_psnr: 32.1898
Epoch 14/40
540/540 [==============================] - 19s 35ms/step - loss: 7.1870e-04 - tf_psnr: 31.5521 - val_loss: 8.0864e-04 - val_tf_psnr: 30.9299
Epoch 15/40
540/540 [==============================] - 19s 35ms/step - loss: 6.8621e-04 - tf_psnr: 31.6766 - val_loss: 5.9929e-04 - val_tf_psnr: 32.2365
Epoch 16/40
540/540 [==============================] - 19s 35ms/step - loss: 7.2689e-04 - tf_psnr: 31.5204 - val_loss: 8.9305e-04 - val_tf_psnr: 30.5157
Epoch 17/40
540/540 [==============================] - 19s 35ms/step - loss: 7.5388e-04 - tf_psnr: 31.3420 - val_loss: 6.9770e-04 - val_tf_psnr: 31.5760
Epoch 18/40
540/540 [==============================] - 19s 35ms/step - loss: 6.7351e-04 - tf_psnr: 31.8044 - val_loss: 6.8552e-04 - val_tf_psnr: 31.6556
Epoch 19/40
540/540 [==============================] - 19s 35ms/step - loss: 0.0038 - tf_psnr: 27.8579 - val_loss: 0.1067 - val_tf_psnr: 9.7576
Epoch 20/40
540/540 [==============================] - 19s 35ms/step - loss: 0.0100 - tf_psnr: 23.9637 - val_loss: 7.0196 - val_tf_psnr: -8.4257
Epoch 21/40
540/540 [==============================] - 19s 35ms/step - loss: 0.0011 - tf_psnr: 29.5019 - val_loss: 0.0052 - val_tf_psnr: 24.7202
Epoch 22/40
540/540 [==============================] - 19s 35ms/step - loss: 9.1100e-04 - tf_psnr: 30.4118 - val_loss: 0.0020 - val_tf_psnr: 27.4877
Epoch 23/40
540/540 [==============================] - 19s 35ms/step - loss: 8.4341e-04 - tf_psnr: 30.7515 - val_loss: 8.7340e-04 - val_tf_psnr: 30.6866
Epoch 24/40
540/540 [==============================] - 19s 35ms/step - loss: 7.9827e-04 - tf_psnr: 30.9913 - val_loss: 8.9695e-04 - val_tf_psnr: 30.4866
Epoch 25/40
540/540 [==============================] - 19s 35ms/step - loss: 7.7264e-04 - tf_psnr: 31.1276 - val_loss: 6.9286e-04 - val_tf_psnr: 31.6229
Epoch 26/40
540/540 [==============================] - 19s 35ms/step - loss: 7.3970e-04 - tf_psnr: 31.3195 - val_loss: 6.8380e-04 - val_tf_psnr: 31.6801
Epoch 27/40
540/540 [==============================] - 19s 35ms/step - loss: 7.3843e-04 - tf_psnr: 31.3281 - val_loss: 7.0879e-04 - val_tf_psnr: 31.5225
Epoch 28/40
540/540 [==============================] - 19s 35ms/step - loss: 7.0559e-04 - tf_psnr: 31.5243 - val_loss: 6.3572e-04 - val_tf_psnr: 32.0017
Epoch 29/40
540/540 [==============================] - 19s 35ms/step - loss: 6.9338e-04 - tf_psnr: 31.6012 - val_loss: 6.1666e-04 - val_tf_psnr: 32.1169
Epoch 30/40
540/540 [==============================] - 19s 35ms/step - loss: 6.7773e-04 - tf_psnr: 31.6996 - val_loss: 6.1554e-04 - val_tf_psnr: 32.1326
Epoch 31/40
540/540 [==============================] - 19s 35ms/step - loss: 6.5468e-04 - tf_psnr: 31.8491 - val_loss: 6.1902e-04 - val_tf_psnr: 32.1035
Epoch 32/40
540/540 [==============================] - 19s 35ms/step - loss: 6.4296e-04 - tf_psnr: 31.9295 - val_loss: 5.8644e-04 - val_tf_psnr: 32.3353
Epoch 33/40
540/540 [==============================] - 19s 35ms/step - loss: 6.2382e-04 - tf_psnr: 32.0584 - val_loss: 5.7848e-04 - val_tf_psnr: 32.3948
Epoch 34/40
540/540 [==============================] - 19s 35ms/step - loss: 6.3895e-04 - tf_psnr: 31.9555 - val_loss: 7.9203e-04 - val_tf_psnr: 31.0150
Epoch 35/40
540/540 [==============================] - 19s 35ms/step - loss: 6.1542e-04 - tf_psnr: 32.1189 - val_loss: 5.4819e-04 - val_tf_psnr: 32.6290
Epoch 36/40
540/540 [==============================] - 19s 35ms/step - loss: 5.9707e-04 - tf_psnr: 32.2535 - val_loss: 5.3602e-04 - val_tf_psnr: 32.7234
Epoch 37/40
540/540 [==============================] - 19s 35ms/step - loss: 5.8027e-04 - tf_psnr: 32.3750 - val_loss: 5.1697e-04 - val_tf_psnr: 32.8870
Epoch 38/40
540/540 [==============================] - 19s 35ms/step - loss: 5.6443e-04 - tf_psnr: 32.4924 - val_loss: 5.1814e-04 - val_tf_psnr: 32.8796
Epoch 39/40
540/540 [==============================] - 19s 35ms/step - loss: 5.7418e-04 - tf_psnr: 32.4189 - val_loss: 5.5014e-04 - val_tf_psnr: 32.6038
Epoch 40/40
540/540 [==============================] - 19s 35ms/step - loss: 5.4728e-04 - tf_psnr: 32.6273 - val_loss: 5.1227e-04 - val_tf_psnr: 32.9183
Save weights for the future reuse.
ddnet_model.save_weights("/content/drive/My Drive/Colab Notebooks/dd_net/model.h5")
Plotting PSNR values to see the trend.
import matplotlib.pyplot as plt
acc = history.history['tf_psnr']
val_acc = history.history['val_tf_psnr']
loss = history.history['loss']
val_loss = history.history['val_loss']
epochs = range(len(acc))
plt.plot(epochs, acc, 'r', label='Training accuracy')
plt.plot(epochs, val_acc, 'b', label='Validation accuracy')
plt.title('Training and validation accuracy')
plt.legend(loc=0)
plt.figure()
plt.show()
<Figure size 432x288 with 0 Axes>
reconstructed = ddnet_model.predict(sparseview_60_test)
psnr = cal_psnr(ndct_test, reconstructed)
print("psnr 40 epochs.....",psnr)
psnr 40 epochs..... 29.894562583454164
psnr: 32.38
from PIL import Image
a = reconstructed[0].reshape(512, 512)
scalef = np.amax(a)
a = np.clip(255 * a/scalef, 0, 255).astype('uint8')
#result = Image.fromarray((a * 255).astype(np.uint8))
result = Image.fromarray((a).astype(np.uint8))
# result.save('unet_15_600_0.png')
result.save('/content/drive/My Drive/Colab Notebooks/dd_net/reconstructed_ddnet_0.png')
a = reconstructed[99].reshape(512, 512)
scalef = np.amax(a)
a = np.clip(255 * a/scalef, 0, 255).astype('uint8')
#result = Image.fromarray((a * 255).astype(np.uint8))
result = Image.fromarray((a).astype(np.uint8))
# result.save('unet_15_600_99.png')
result.save('/content/drive/My Drive/Colab Notebooks/dd_net/reconstructed_ddnet_99.png')
a = sparseview_60_test[0].reshape(512, 512)
scalef = np.amax(a)
a = np.clip(255 * a/scalef, 0, 255).astype('uint8')
#result = Image.fromarray((a * 255).astype(np.uint8))
result = Image.fromarray((a).astype(np.uint8))
# result.save('unet_15_600_0.png')
result.save('/content/drive/My Drive/Colab Notebooks/dd_net/sparseview_60_test_ddnet_0.png')
a = sparseview_60_test[99].reshape(512, 512)
scalef = np.amax(a)
a = np.clip(255 * a/scalef, 0, 255).astype('uint8')
#result = Image.fromarray((a * 255).astype(np.uint8))
result = Image.fromarray((a).astype(np.uint8))
# result.save('unet_15_600_99.png')
result.save('/content/drive/My Drive/Colab Notebooks/dd_net/sparseview_60_test_ddnet_99.png')
# Plot training & validation accuracy values
plt.plot(history.history['tf_psnr'])
plt.plot(history.history['val_tf_psnr'])
plt.title('Model accuracy')
plt.ylabel('Accuracy')
plt.xlabel('Epoch')
plt.legend(['Train', 'Test'], loc='upper left')
plt.show()
# Plot training & validation loss values
plt.plot(history.history['loss'])
plt.plot(history.history['val_loss'])
plt.title('Model loss')
plt.ylabel('Loss')
plt.xlabel('Epoch')
plt.legend(['Train', 'Test'], loc='upper left')
plt.show()
img=mpimg.imread('/content/drive/My Drive/Colab Notebooks/dd_net/sparseview_60_test_ddnet_0.png')
imgplot = plt.imshow(img)
plt.show()
img=mpimg.imread('/content/drive/My Drive/Colab Notebooks/dd_net/reconstructed_ddnet_0.png')
imgplot = plt.imshow(img)
plt.show()
img=mpimg.imread('/content/drive/My Drive/Colab Notebooks/dd_net/sparseview_60_test_ddnet_99.png')
imgplot = plt.imshow(img)
plt.show()
img=mpimg.imread('/content/drive/My Drive/Colab Notebooks/dd_net/reconstructed_ddnet_99.png')
imgplot = plt.imshow(img)
plt.show()
