Medical_image_denoising_SSIM.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import cv2\n",
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "from skimage.util import random_noise"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(322, 4096)\n"
     ]
    }
   ],
   "source": [
    "# Flattening the images together\n",
    "path = 'PATH_TO_DATA_FOLDER'\n",
    "img_single = [] # storing all the flattened images.\n",
    "train_y = [] # storing the original images without any noise as ground truths.\n",
    "test_y = []\n",
    "\n",
    "for i,file in enumerate(sorted(os.listdir(path))):\n",
    "    try:\n",
    "        image = cv2.imread(os.path.join(path, file), 0) # grayscale\n",
    "        img_resized = cv2.resize(image, (64,64))\n",
    "        image_flat = img_resized.flatten()\n",
    "        img_single.append(image_flat)\n",
    "        if(i<=30): #10% of 300 images in test data\n",
    "            test_y.append(img_resized)\n",
    "        else:\n",
    "            train_y.append(img_resized)\n",
    "    except Exception as e:\n",
    "        continue\n",
    "\n",
    "img_single_arr = np.asarray(img_single)\n",
    "print(img_single_arr.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(292, 64, 64, 1)\n",
      "(30, 64, 64, 1)\n"
     ]
    }
   ],
   "source": [
    "train_y_arr = np.asarray(train_y)/255.0 # All the original and noised images should have values = [0,1]\n",
    "test_y_arr = np.asarray(test_y)/255.0\n",
    "train_y_arr = train_y_arr[:,:,:,np.newaxis].astype('float32') # making (batch, width, height, channel) form\n",
    "test_y_arr = test_y_arr[:,:,:,np.newaxis].astype('float32')\n",
    "print(train_y_arr.shape)\n",
    "print(test_y_arr.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(322, 4096)\n"
     ]
    }
   ],
   "source": [
    "# Adding salt and pepper noise with 10% amount\n",
    "noisy_img_scaled = random_noise(img_single_arr, mode='s&p',  amount=0.1)\n",
    "print(noisy_img_scaled.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(292, 64, 64, 1)\n",
      "(30, 64, 64, 1)\n"
     ]
    }
   ],
   "source": [
    "# Reshaping back into individual images\n",
    "train_x = [] #storining the noised images for training the AE\n",
    "test_x = []\n",
    "for x in range(0, noisy_img_scaled.shape[0]):\n",
    "    img_arr = noisy_img_scaled[x, :].reshape(64,64)\n",
    "    if(x<30): #10% of 300 images for test data\n",
    "        test_x.append(img_arr)\n",
    "    else:\n",
    "        train_x.append(img_arr)\n",
    "\n",
    "train_x_arr = np.asarray(train_x)\n",
    "train_x_arr = train_x_arr[:,:,:,np.newaxis] # making (batch, width, height, channel) form\n",
    "print(train_x_arr.shape)\n",
    "test_x_arr = np.asarray(test_x)\n",
    "test_x_arr = test_x_arr[:,:,:,np.newaxis]\n",
    "print(test_x_arr.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Defining the Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.models import Model\n",
    "from tensorflow.keras import layers\n",
    "from tensorflow.keras.utils import plot_model\n",
    "from skimage.metrics import structural_similarity as ssim"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# model architecture\n",
    "input = layers.Input(shape=(64,64,1))\n",
    "down = layers.Conv2D(64, (3,3), activation='relu', padding='same')(input)\n",
    "down = layers.MaxPooling2D(pool_size=(2,2))(down)\n",
    "down = layers.Conv2D(64, (3,3), activation='relu', padding='same')(down)\n",
    "down = layers.MaxPooling2D(pool_size=(2,2))(down)\n",
    "down = layers.Conv2D(64, (3,3), activation='relu', padding='same')(down)\n",
    "up = layers.UpSampling2D(2)(down)\n",
    "up = layers.Conv2D(64, (3,3), activation='relu', padding='same')(up)\n",
    "up = layers.UpSampling2D(2)(up)\n",
    "output = layers.Conv2D(1, (3,3), activation='sigmoid', padding='same')(up)\n",
    "\n",
    "model = Model(inputs=input, outputs=output)\n",
    "\n",
    "plot_model(model, to_file='model_plot.png', show_shapes=True, show_layer_names=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_1 (InputLayer)         [(None, 64, 64, 1)]       0         \n",
      "_________________________________________________________________\n",
      "conv2d (Conv2D)              (None, 64, 64, 64)        640       \n",
      "_________________________________________________________________\n",
      "max_pooling2d (MaxPooling2D) (None, 32, 32, 64)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_1 (Conv2D)            (None, 32, 32, 64)        36928     \n",
      "_________________________________________________________________\n",
      "max_pooling2d_1 (MaxPooling2 (None, 16, 16, 64)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_2 (Conv2D)            (None, 16, 16, 64)        36928     \n",
      "_________________________________________________________________\n",
      "up_sampling2d (UpSampling2D) (None, 32, 32, 64)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_3 (Conv2D)            (None, 32, 32, 64)        36928     \n",
      "_________________________________________________________________\n",
      "up_sampling2d_1 (UpSampling2 (None, 64, 64, 64)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_4 (Conv2D)            (None, 64, 64, 1)         577       \n",
      "=================================================================\n",
      "Total params: 112,001\n",
      "Trainable params: 112,001\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Structural Similarity Index as a loss function\n",
    "\n",
    "def loss_fn(y_true, y_pred):\n",
    "    return(1 - tf.reduce_mean(tf.image.ssim(y_true, y_pred, 1.0)))\n",
    "    #max_val=1 because we have image values=[0,1]\n",
    "    # Returning=(1-SSIM) as we want to minimize the less similar images"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Time based learning rate scheduling\n",
    "\n",
    "epoch=100\n",
    "# learning_rate=0.01\n",
    "# decay_rate=learning_rate/epoch\n",
    "# momentum=0.8\n",
    "# sgd=SGD(lr=learning_rate,momentum=momentum,decay=decay_rate)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Early stopping callback\n",
    "callback = tf.keras.callbacks.EarlyStopping(monitor='val_loss', patience=5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "# compiling the model \n",
    "model.compile(loss=loss_fn, optimizer='rmsprop', run_eagerly=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n",
      "27/27 [==============================] - 18s 661ms/step - loss: 0.6113 - val_loss: 0.3805\n",
      "Epoch 2/100\n",
      "27/27 [==============================] - 18s 658ms/step - loss: 0.2751 - val_loss: 0.2282\n",
      "Epoch 3/100\n",
      "27/27 [==============================] - 17s 639ms/step - loss: 0.2362 - val_loss: 0.1669\n",
      "Epoch 4/100\n",
      "27/27 [==============================] - 17s 642ms/step - loss: 0.2156 - val_loss: 0.1731\n",
      "Epoch 5/100\n",
      "27/27 [==============================] - 17s 648ms/step - loss: 0.2015 - val_loss: 0.1621\n",
      "Epoch 6/100\n",
      "27/27 [==============================] - 17s 638ms/step - loss: 0.1830 - val_loss: 0.1582\n",
      "Epoch 7/100\n",
      "27/27 [==============================] - 17s 638ms/step - loss: 0.1690 - val_loss: 0.1770\n",
      "Epoch 8/100\n",
      "27/27 [==============================] - 17s 640ms/step - loss: 0.1615 - val_loss: 0.1510\n",
      "Epoch 9/100\n",
      "27/27 [==============================] - 17s 641ms/step - loss: 0.1570 - val_loss: 0.2288\n",
      "Epoch 10/100\n",
      "27/27 [==============================] - 17s 640ms/step - loss: 0.1615 - val_loss: 0.1357\n",
      "Epoch 11/100\n",
      "27/27 [==============================] - 17s 644ms/step - loss: 0.1437 - val_loss: 0.1757\n",
      "Epoch 12/100\n",
      "27/27 [==============================] - 17s 646ms/step - loss: 0.1417 - val_loss: 0.1214\n",
      "Epoch 13/100\n",
      "27/27 [==============================] - 17s 639ms/step - loss: 0.1307 - val_loss: 0.1219\n",
      "Epoch 14/100\n",
      "27/27 [==============================] - 17s 646ms/step - loss: 0.1259 - val_loss: 0.1118\n",
      "Epoch 15/100\n",
      "27/27 [==============================] - 17s 642ms/step - loss: 0.1276 - val_loss: 0.1332\n",
      "Epoch 16/100\n",
      "27/27 [==============================] - 17s 641ms/step - loss: 0.1263 - val_loss: 0.1271\n",
      "Epoch 17/100\n",
      "27/27 [==============================] - 17s 643ms/step - loss: 0.1198 - val_loss: 0.1324\n",
      "Epoch 18/100\n",
      "27/27 [==============================] - 17s 638ms/step - loss: 0.1191 - val_loss: 0.1018\n",
      "Epoch 19/100\n",
      "27/27 [==============================] - 17s 646ms/step - loss: 0.1116 - val_loss: 0.1175\n",
      "Epoch 20/100\n",
      "27/27 [==============================] - 17s 643ms/step - loss: 0.1177 - val_loss: 0.1128\n",
      "Epoch 21/100\n",
      "27/27 [==============================] - 17s 646ms/step - loss: 0.1104 - val_loss: 0.1194\n",
      "Epoch 22/100\n",
      "27/27 [==============================] - 17s 646ms/step - loss: 0.1073 - val_loss: 0.0977\n",
      "Epoch 23/100\n",
      "27/27 [==============================] - 17s 641ms/step - loss: 0.1021 - val_loss: 0.1301\n",
      "Epoch 24/100\n",
      "27/27 [==============================] - 17s 644ms/step - loss: 0.1043 - val_loss: 0.1136\n",
      "Epoch 25/100\n",
      "27/27 [==============================] - 17s 641ms/step - loss: 0.1022 - val_loss: 0.0943\n",
      "Epoch 26/100\n",
      "27/27 [==============================] - 17s 640ms/step - loss: 0.0958 - val_loss: 0.0975\n",
      "Epoch 27/100\n",
      "27/27 [==============================] - 17s 641ms/step - loss: 0.0959 - val_loss: 0.0900\n",
      "Epoch 28/100\n",
      "27/27 [==============================] - 17s 643ms/step - loss: 0.0964 - val_loss: 0.1065\n",
      "Epoch 29/100\n",
      "27/27 [==============================] - 17s 645ms/step - loss: 0.0977 - val_loss: 0.0984\n",
      "Epoch 30/100\n",
      "27/27 [==============================] - 17s 644ms/step - loss: 0.1002 - val_loss: 0.1100\n",
      "Epoch 31/100\n",
      "27/27 [==============================] - 17s 642ms/step - loss: 0.0982 - val_loss: 0.0994\n",
      "Epoch 32/100\n",
      "27/27 [==============================] - 17s 640ms/step - loss: 0.0901 - val_loss: 0.1075\n"
     ]
    }
   ],
   "source": [
    "model_history=model.fit(train_x_arr, train_y_arr, batch_size = 10, epochs = epoch,\n",
    "                        validation_split = 0.1, shuffle = True, callbacks = [callback])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Plotting results and predicted images"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plotting the varoius metrics\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "pd.DataFrame(model_history.history).plot(figsize=(8,5))\n",
    "plt.grid(True)\n",
    "plt.gca().set_ylim(0,0.5)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1/1 [==============================] - 0s 418ms/step - loss: 0.1069\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.10688400268554688"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Testing the model\n",
    "model.evaluate(test_x_arr, test_y_arr)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Noised and Denoised results:**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(64, 64, 1)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7f6d6b7d8e10>"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "test_img = test_x_arr[0]\n",
    "print(test_img.shape)\n",
    "plt.imshow(test_img)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1, 64, 64, 1)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7f6d6b75acf8>"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "test_img_axis = test_img[np.newaxis,:,:,:]\n",
    "# print(test_img_axis.shape)\n",
    "test_img_pred = model.predict(test_img_axis)\n",
    "\n",
    "print(test_img_pred.shape)\n",
    "plt.imshow(np.squeeze(test_img_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Conclusion\n",
    "This code is an implementation of the paper - ***Medical image denoising using convolutional\n",
    "denoising autoencoders*** (https://arxiv.org/pdf/1608.04667.pdf).\n",
    "\n",
    "The model performs satisfactorily well given the small amount of data (less than 300 for training) and a very simple model architecture. This proves the point of the paper that we can use AutoEncoders for denoising even with very less sample of data.\n",
    "The data used for this experiment was - **MIAS MiniMammographic Database**, (REFERENCE: *J Suckling et al (1994) \"The Mammographic Image Analysis Society Digital Mammogram Database\" Exerpta Medica. International Congress Series 1069 pp375-378.*), containing 322 images in total. I used 30 images for test and out of 292 remaining, I made a validation split of 10% while training.\n",
    "\n",
    "The model evaluation gives a loss of 0.106, ie Loss = 1-SSIM => SSIM = 1-0.106 = **0.894** which is quite well.\n",
    "From further here, we can improve this work to try using MSE loss, different noises, deeper architectures with skip connections, more data, etc."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.9"
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}