SISRIA-Model.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Single Image Super Resolution Using Deep Convolutional Network"
   ]
  },
  {
   "cell_type": "raw",
   "metadata": {},
   "source": [
    "Md Kamrul Hasan\n",
    "Yujie"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Initialising all the pacage and necessary library"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Conv2D\n",
    "from keras.optimizers import SGD, adam\n",
    "import psnr\n",
    "import numpy\n",
    "import math\n",
    "import os\n",
    "import cv2\n",
    "import h5py\n",
    "import numpy\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# PSNR Function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "def psnr(target, ref):\n",
    "    # assume RGB image\n",
    "    target_data = numpy.array(target, dtype=float)\n",
    "    ref_data = numpy.array(ref, dtype=float)\n",
    "\n",
    "    diff = ref_data - target_data\n",
    "    diff = diff.flatten('C')\n",
    "\n",
    "    rmse = math.sqrt(numpy.mean(diff ** 2.))\n",
    "\n",
    "    return 20 * math.log10(255. / rmse)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Readin Training Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def read_training_data(file):\n",
    "    with h5py.File(file, 'r') as hf:\n",
    "        data = numpy.array(hf.get('data'))\n",
    "        label = numpy.array(hf.get('label'))\n",
    "        train_data = numpy.transpose(data, (0, 2, 3, 1))\n",
    "        train_label = numpy.transpose(label, (0, 2, 3, 1))\n",
    "        return train_data, train_label"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Designing the CNN Model architecture "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "def model():\n",
    "    SRCNNIA = Sequential()\n",
    "    SRCNNIA.add(Conv2D(64,(9,9),padding='same',input_shape = (None,None,1)))\n",
    "    SRCNNIA.add(Activation('relu')) \n",
    "    SRCNNIA.add(Conv2D(32,(1,1),padding='same'))\n",
    "    SRCNNA.add(Activation('relu'))\n",
    "    SRCNNIA.add(Conv2D(1,(5,5),padding='same'))\n",
    "    Adam = adam(lr=0.001)\n",
    "    SRCNNIA.compile(optimizer=Adam, loss='mean_squared_error', metrics=['mean_squared_error'])\n",
    "    \n",
    "    return SRCNNIA\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Predicting the HR Image from LR image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [],
   "source": [
    "def predict():\n",
    "    srcnn_model = model()\n",
    "    srcnn_model.load_weights(\"m_model_adam30_64.h5\")\n",
    "    IMG_NAMEO =\"./Test/Set14/WTruckHr.JPG\"\n",
    "    IMG_NAME = \"./Test/Set14/WTruckLr.jpg\"\n",
    "    INPUT_NAME = \"inputWR.jpg\"\n",
    "    OUTPUT_NAME = \"pre_adam30cecropWRN.jpg\"\n",
    "    \n",
    "\n",
    "    import cv2\n",
    "    imgo = cv2.imread(IMG_NAMEO)\n",
    "    img = cv2.imread(IMG_NAME)\n",
    "    shape = img.shape\n",
    "    #img = cv2.resize(img, (int(shape[1] / 2), int(shape[0] / 2)))\n",
    "    \n",
    "    img = cv2.resize(img, (shape[1]*2, shape[0]*2), cv2.INTER_CUBIC)\n",
    "    cv2.imwrite(INPUT_NAME, img)\n",
    "    img = cv2.cvtColor(img, cv2.COLOR_BGR2YCrCb)\n",
    "    Y = numpy.zeros((1, img.shape[0], img.shape[1], 1))\n",
    "    Y[0, :, :, 0] = img[:, :, 0]\n",
    "    pre = srcnn_model.predict(Y, batch_size=1)\n",
    "    pre[pre[:] > 255] = 255\n",
    "    pre[pre[:] < 0] = 0\n",
    "    pre = pre.astype(numpy.uint8)\n",
    "    img[6: -6, 6: -6, 0] = pre[0, :, :, 0]\n",
    "    img = cv2.cvtColor(img, cv2.COLOR_YCrCb2BGR)\n",
    "    cv2.imwrite(OUTPUT_NAME, img)\n",
    "\n",
    "    # psnr calculation:\n",
    "    im1 = cv2.imread(IMG_NAME, cv2.IMREAD_COLOR)\n",
    "    im1 = cv2.cvtColor(im1, cv2.COLOR_BGR2YCrCb)[6: -6, 6: -6, 0]\n",
    "    im2 = cv2.imread(INPUT_NAME, cv2.IMREAD_COLOR)\n",
    "    im2 = cv2.cvtColor(im2, cv2.COLOR_BGR2YCrCb)[6: -6, 6: -6, 0]\n",
    "    im3 = cv2.imread(OUTPUT_NAME, cv2.IMREAD_COLOR)\n",
    "    im3 = cv2.cvtColor(im3, cv2.COLOR_BGR2YCrCb)[6: -6, 6: -6, 0]\n",
    "    im4 = cv2.imread(IMG_NAMEO, cv2.IMREAD_COLOR)\n",
    "    im4 = cv2.cvtColor(im4, cv2.COLOR_BGR2YCrCb)[6: -6, 6: -6, 0]\n",
    "\n",
    "    print(\"bicubic:\")\n",
    "    print(psnr(im1, im2))\n",
    "    print(\"SRCNN:\")\n",
    "    print(psnr(im3, im4))\n",
    "    print(history.history.keys())\n",
    "    plt.figure(1)\n",
    "    plt.plot(history.history['loss'])\n",
    "    plt.title('mean_squared_error')\n",
    "    plt.ylabel('loss')\n",
    "    plt.xlabel('epoch')\n",
    "    plt.show()\n",
    "    plt.savefig(\"m_model_adam30_ce.png\")\n",
    "    \n",
    "    print(history.model.layers[1].output)\n",
    "    \n",
    "   # psnr = compare_psnr(im1,im3)\n",
    "   # print(ssim)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Running the model and saving the model result with .h5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\hasan\\Anaconda3\\envs\\DL\\lib\\site-packages\\ipykernel_launcher.py:4: UserWarning: Update your `Conv2D` call to the Keras 2 API: `Conv2D(activation=\"relu\", filters=64, use_bias=True, kernel_size=(9, 9), kernel_initializer=\"he_normal\", input_shape=(None, Non..., padding=\"valid\")`\n",
      "  after removing the cwd from sys.path.\n",
      "C:\\Users\\hasan\\Anaconda3\\envs\\DL\\lib\\site-packages\\ipykernel_launcher.py:6: UserWarning: Update your `Conv2D` call to the Keras 2 API: `Conv2D(activation=\"relu\", filters=32, padding=\"valid\", use_bias=True, kernel_size=(1, 1), kernel_initializer=\"he_normal\")`\n",
      "  \n",
      "C:\\Users\\hasan\\Anaconda3\\envs\\DL\\lib\\site-packages\\ipykernel_launcher.py:8: UserWarning: Update your `Conv2D` call to the Keras 2 API: `Conv2D(activation=\"linear\", filters=1, padding=\"valid\", use_bias=True, kernel_size=(5, 5), kernel_initializer=\"he_normal\")`\n",
      "  \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d_106 (Conv2D)          (None, None, None, 64)    5248      \n",
      "_________________________________________________________________\n",
      "conv2d_107 (Conv2D)          (None, None, None, 32)    2080      \n",
      "_________________________________________________________________\n",
      "conv2d_108 (Conv2D)          (None, None, None, 1)     801       \n",
      "=================================================================\n",
      "Total params: 8,129\n",
      "Trainable params: 8,129\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\hasan\\Anaconda3\\envs\\DL\\lib\\site-packages\\keras\\models.py:981: UserWarning: The `nb_epoch` argument in `fit` has been renamed `epochs`.\n",
      "  warnings.warn('The `nb_epoch` argument in `fit` '\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/30\n",
      "40000/40000 [==============================] - 97s 2ms/step - loss: 0.0043 - mean_squared_error: 0.0043\n",
      "Epoch 2/30\n",
      "40000/40000 [==============================] - 94s 2ms/step - loss: 0.0036 - mean_squared_error: 0.0036\n",
      "Epoch 3/30\n",
      "40000/40000 [==============================] - 94s 2ms/step - loss: 0.0036 - mean_squared_error: 0.0036\n",
      "Epoch 4/30\n",
      "40000/40000 [==============================] - 97s 2ms/step - loss: 0.0035 - mean_squared_error: 0.0035\n",
      "Epoch 5/30\n",
      "40000/40000 [==============================] - 97s 2ms/step - loss: 0.0035 - mean_squared_error: 0.0035\n",
      "Epoch 6/30\n",
      "40000/40000 [==============================] - 97s 2ms/step - loss: 0.0035 - mean_squared_error: 0.0035\n",
      "Epoch 7/30\n",
      "40000/40000 [==============================] - 98s 2ms/step - loss: 0.0035 - mean_squared_error: 0.0035\n",
      "Epoch 8/30\n",
      "40000/40000 [==============================] - 97s 2ms/step - loss: 0.0034 - mean_squared_error: 0.0034\n",
      "Epoch 9/30\n",
      "40000/40000 [==============================] - 98s 2ms/step - loss: 0.0034 - mean_squared_error: 0.0034\n",
      "Epoch 10/30\n",
      "40000/40000 [==============================] - 99s 2ms/step - loss: 0.0034 - mean_squared_error: 0.0034\n",
      "Epoch 11/30\n",
      "40000/40000 [==============================] - 100s 3ms/step - loss: 0.0034 - mean_squared_error: 0.0034\n",
      "Epoch 12/30\n",
      "40000/40000 [==============================] - 104s 3ms/step - loss: 0.0034 - mean_squared_error: 0.0034\n",
      "Epoch 13/30\n",
      "40000/40000 [==============================] - 103s 3ms/step - loss: 0.0035 - mean_squared_error: 0.0035\n",
      "Epoch 14/30\n",
      "40000/40000 [==============================] - 98s 2ms/step - loss: 0.0034 - mean_squared_error: 0.0034\n",
      "Epoch 15/30\n",
      "40000/40000 [==============================] - 97s 2ms/step - loss: 0.0034 - mean_squared_error: 0.0034\n",
      "Epoch 16/30\n",
      "40000/40000 [==============================] - 100s 2ms/step - loss: 0.0034 - mean_squared_error: 0.0034\n",
      "Epoch 17/30\n",
      "40000/40000 [==============================] - 103s 3ms/step - loss: 0.0034 - mean_squared_error: 0.0034\n",
      "Epoch 18/30\n",
      "40000/40000 [==============================] - 100s 3ms/step - loss: 0.0034 - mean_squared_error: 0.0034\n",
      "Epoch 19/30\n",
      "40000/40000 [==============================] - 100s 3ms/step - loss: 0.0034 - mean_squared_error: 0.0034\n",
      "Epoch 20/30\n",
      "40000/40000 [==============================] - 100s 2ms/step - loss: 0.0034 - mean_squared_error: 0.0034\n",
      "Epoch 21/30\n",
      "40000/40000 [==============================] - 99s 2ms/step - loss: 0.0034 - mean_squared_error: 0.0034\n",
      "Epoch 22/30\n",
      "40000/40000 [==============================] - 106s 3ms/step - loss: 0.0034 - mean_squared_error: 0.0034\n",
      "Epoch 23/30\n",
      "40000/40000 [==============================] - 122s 3ms/step - loss: 0.0034 - mean_squared_error: 0.0034\n",
      "Epoch 24/30\n",
      "40000/40000 [==============================] - 101s 3ms/step - loss: 0.0034 - mean_squared_error: 0.0034\n",
      "Epoch 25/30\n",
      "40000/40000 [==============================] - 103s 3ms/step - loss: 0.0034 - mean_squared_error: 0.0034\n",
      "Epoch 26/30\n",
      "40000/40000 [==============================] - 101s 3ms/step - loss: 0.0034 - mean_squared_error: 0.0034\n",
      "Epoch 27/30\n",
      "40000/40000 [==============================] - 101s 3ms/step - loss: 0.0034 - mean_squared_error: 0.0034\n",
      "Epoch 28/30\n",
      "40000/40000 [==============================] - 103s 3ms/step - loss: 0.0034 - mean_squared_error: 0.0034\n",
      "Epoch 29/30\n",
      "40000/40000 [==============================] - 103s 3ms/step - loss: 0.0034 - mean_squared_error: 0.0034\n",
      "Epoch 30/30\n",
      "40000/40000 [==============================] - 103s 3ms/step - loss: 0.0034 - mean_squared_error: 0.0034\n",
      "dict_keys(['mean_squared_error', 'loss'])\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "srcnn_model = model()\n",
    "srcnn_model.summary()\n",
    "data, label = read_training_data(\"./train.h5\")\n",
    "    # srcnn_model.load_weights(\"m_model_adam.h5\")\n",
    "history=srcnn_model.fit(data, label, batch_size=128, nb_epoch=30)\n",
    "srcnn_model.save_weights(\"m_model_adam30_64.h5\")\n",
    "print(history.history.keys())\n",
    "plt.figure(1)\n",
    "plt.plot(history.history['loss'])\n",
    "plt.title('model loss')\n",
    "plt.ylabel('loss')\n",
    "plt.xlabel('epoch')\n",
    "plt.show()\n",
    "plt.savefig(\"m_model_adam30_ce.png\")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Final rediction from main function callding the predict() function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\hasan\\Anaconda3\\envs\\DL\\lib\\site-packages\\ipykernel_launcher.py:4: UserWarning: Update your `Conv2D` call to the Keras 2 API: `Conv2D(activation=\"relu\", filters=64, use_bias=True, kernel_size=(9, 9), kernel_initializer=\"he_normal\", input_shape=(None, Non..., padding=\"valid\")`\n",
      "  after removing the cwd from sys.path.\n",
      "C:\\Users\\hasan\\Anaconda3\\envs\\DL\\lib\\site-packages\\ipykernel_launcher.py:6: UserWarning: Update your `Conv2D` call to the Keras 2 API: `Conv2D(activation=\"relu\", filters=32, padding=\"valid\", use_bias=True, kernel_size=(1, 1), kernel_initializer=\"he_normal\")`\n",
      "  \n",
      "C:\\Users\\hasan\\Anaconda3\\envs\\DL\\lib\\site-packages\\ipykernel_launcher.py:8: UserWarning: Update your `Conv2D` call to the Keras 2 API: `Conv2D(activation=\"linear\", filters=1, padding=\"valid\", use_bias=True, kernel_size=(5, 5), kernel_initializer=\"he_normal\")`\n",
      "  \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "bicubic:\n",
      "SRCNN:\n",
      "18.550807961693828\n",
      "dict_keys(['mean_squared_error', 'loss'])\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tensor(\"conv2d_107/Relu:0\", shape=(?, ?, ?, 32), dtype=float32)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "if __name__ == \"__main__\":\n",
    "    #train()\n",
    "    predict()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 30 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "weight_conv2d_1 = history.model.layers[0].get_weights()[0][:,:,0,:]\n",
    " \n",
    "col_size = 6\n",
    "row_size = 5\n",
    "filter_index = 0\n",
    "fig, ax = plt.subplots(row_size, col_size, figsize=(12,8))\n",
    "for row in range(0,row_size): \n",
    "  for col in range(0,col_size):\n",
    "    ax[row][col].imshow(weight_conv2d_1[:,:,filter_index],cmap=\"gray\")\n",
    "    filter_index += 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.5.5"
  },
  "widgets": {
   "state": {},
   "version": "1.1.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}