CNN for Image Classification.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Develop Convolutional Neural Network Models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# import packages\n",
    "from numpy import unique\n",
    "from numpy import argmax\n",
    "from tensorflow.keras.datasets.mnist import load_data\n",
    "from tensorflow.keras import Sequential\n",
    "from tensorflow.keras.layers import Dense\n",
    "from tensorflow.keras.layers import Conv2D\n",
    "from tensorflow.keras.layers import MaxPooling2D\n",
    "from tensorflow.keras.layers import Flatten\n",
    "from tensorflow.keras.layers import Dropout\n",
    "from matplotlib import pyplot \n",
    "import warnings \n",
    "warnings.filterwarnings(\"ignore\") "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "A popular image classification task is the MNIST handwritten digit classification. It involves tens of thousands of handwritten digits that must be classified as a number between 0 and 9.\n",
    "\n",
    "The tf.keras API provides a convenience function to download and load this dataset directly."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# load dataset\n",
    "(x_train, y_train), (x_test, y_test) = load_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train: X=(60000, 28, 28), y=(60000,)\n",
      "Test: X=(10000, 28, 28), y=(10000,)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 25 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# summarize loaded dataset\n",
    "print('Train: X=%s, y=%s' % (x_train.shape, y_train.shape))\n",
    "print('Test: X=%s, y=%s' % (x_test.shape, y_test.shape))\n",
    "# plot first few images\n",
    "for i in range(25):\n",
    "    # define subplot\n",
    "    pyplot.subplot(5, 5, i+1)\n",
    "    # plot raw pixel data\n",
    "    pyplot.imshow(x_train[i], cmap=pyplot.get_cmap('gray'))\n",
    "# show the figure\n",
    "pyplot.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can train a CNN model to classify the images in the MNIST dataset.\n",
    "\n",
    "Note that the images are arrays of grayscale pixel data; therefore, we must add a channel dimension to the data before we can use the images as input to the model. The reason is that CNN models expect images in a channels-last format, that is each example to the network has the dimensions [rows, columns, channels], where channels represent the color channels of the image data."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "It is also a good idea to scale the pixel values from the default range of 0-255 to 0-1 when training a CNN."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# reshape data to have a single channel\n",
    "x_train = x_train.reshape((x_train.shape[0], x_train.shape[1], x_train.shape[2], 1))\n",
    "x_test = x_test.reshape((x_test.shape[0], x_test.shape[1], x_test.shape[2], 1))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(28, 28, 1) 10\n"
     ]
    }
   ],
   "source": [
    "# determine the shape of the input images\n",
    "in_shape = x_train.shape[1:]\n",
    "# determine the number of classes\n",
    "n_classes = len(unique(y_train))\n",
    "print(in_shape, n_classes)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# normalize pixel values\n",
    "x_train = x_train.astype('float32') / 255.0\n",
    "x_test = x_test.astype('float32') / 255.0\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# define model\n",
    "model = Sequential()\n",
    "model.add(Conv2D(32, (3,3), activation='relu', kernel_initializer='he_uniform', input_shape=in_shape))\n",
    "model.add(MaxPooling2D((2, 2)))\n",
    "model.add(Flatten())\n",
    "model.add(Dense(100, activation='relu', kernel_initializer='he_uniform'))\n",
    "model.add(Dropout(0.5))\n",
    "model.add(Dense(n_classes, activation='softmax'))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "# define loss and optimizer\n",
    "model.compile(optimizer='adam', loss='sparse_categorical_crossentropy', metrics=['accuracy'])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 48000 samples, validate on 12000 samples\n",
      "Epoch 1/10\n",
      "48000/48000 - 73s - loss: 0.3483 - accuracy: 0.8936 - val_loss: 0.0964 - val_accuracy: 0.9726\n",
      "Epoch 2/10\n",
      "48000/48000 - 72s - loss: 0.1414 - accuracy: 0.9590 - val_loss: 0.0658 - val_accuracy: 0.9818\n",
      "Epoch 3/10\n",
      "48000/48000 - 70s - loss: 0.1009 - accuracy: 0.9702 - val_loss: 0.0590 - val_accuracy: 0.9832\n",
      "Epoch 4/10\n",
      "48000/48000 - 70s - loss: 0.0810 - accuracy: 0.9751 - val_loss: 0.0581 - val_accuracy: 0.9833\n",
      "Epoch 5/10\n",
      "48000/48000 - 72s - loss: 0.0699 - accuracy: 0.9785 - val_loss: 0.0511 - val_accuracy: 0.9843\n",
      "Epoch 6/10\n",
      "48000/48000 - 70s - loss: 0.0608 - accuracy: 0.9814 - val_loss: 0.0530 - val_accuracy: 0.9843\n",
      "Epoch 7/10\n",
      "48000/48000 - 70s - loss: 0.0553 - accuracy: 0.9832 - val_loss: 0.0536 - val_accuracy: 0.9848\n",
      "Epoch 8/10\n",
      "48000/48000 - 66s - loss: 0.0502 - accuracy: 0.9836 - val_loss: 0.0504 - val_accuracy: 0.9848\n",
      "Epoch 9/10\n",
      "48000/48000 - 67s - loss: 0.0442 - accuracy: 0.9852 - val_loss: 0.0505 - val_accuracy: 0.9858\n",
      "Epoch 10/10\n",
      "48000/48000 - 67s - loss: 0.0403 - accuracy: 0.9865 - val_loss: 0.0505 - val_accuracy: 0.9865\n"
     ]
    }
   ],
   "source": [
    "# fit the model\n",
    "history = model.fit(x_train, y_train, epochs=10, batch_size=128, validation_split=0.2, verbose=2)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "# plot learning curves\n",
    "pyplot.title('Learning Curves')\n",
    "pyplot.xlabel('Epoch')\n",
    "pyplot.ylabel('loss')\n",
    "pyplot.plot(history.history['loss'], label='train')\n",
    "pyplot.plot(history.history['val_loss'], label='val')\n",
    "pyplot.legend()\n",
    "pyplot.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.986\n"
     ]
    }
   ],
   "source": [
    "# evaluate the model\n",
    "loss, acc = model.evaluate(x_test, y_test, verbose=0)\n",
    "print('Accuracy: %.3f' % acc)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Predicted: class=5\n"
     ]
    }
   ],
   "source": [
    "# make a prediction\n",
    "image = x_train[0]\n",
    "yhat = model.predict([[image]])\n",
    "print('Predicted: class=%d' % argmax(yhat))"
   ]
  }
 ],
 "metadata": {
  "hide_input": false,
  "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": "3.7.3"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}