rs_baseline_v2.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-04-27 17:48:34.368490: W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libcudart.so.11.0'; dlerror: libcudart.so.11.0: cannot open shared object file: No such file or directory\n",
      "2022-04-27 17:48:34.368546: I tensorflow/stream_executor/cuda/cudart_stub.cc:29] Ignore above cudart dlerror if you do not have a GPU set up on your machine.\n"
     ]
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "\n",
    "import os\n",
    "os.environ['CUDA_VISIBLE_DEVICES'] = \"0\"\n",
    "\n",
    "import numpy as np\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "import keras\n",
    "\n",
    "from keras.applications.efficientnet import preprocess_input\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.utils import to_categorical\n",
    "from keras.callbacks import ModelCheckpoint, ReduceLROnPlateau\n",
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "from keras.layers import Dense, Dropout, Flatten, Activation, Conv2D, MaxPooling2D, BatchNormalization\n",
    "from keras.preprocessing import image\n",
    "\n",
    "from sklearn.metrics import confusion_matrix, ConfusionMatrixDisplay\n",
    "from sklearn.metrics import precision_score, recall_score, accuracy_score\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from sklearn.utils.class_weight import compute_class_weight"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load Dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Read the test and training data into variables. Images are already processed here, i.e. correctly sized, converted to numpy arrays (..)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['agricultural', 'airplane', 'baseballdiamond', 'beach', 'buildings', 'chaparral', 'denseresidential', 'forest', 'freeway', 'golfcourse', 'harbor', 'intersection', 'mediumresidential', 'mobilehomepark', 'overpass', 'parkinglot', 'river', 'runway', 'sparseresidential', 'storagetanks', 'tenniscourt']\n"
     ]
    }
   ],
   "source": [
    "train_dir = './ds/train/'\n",
    "test_dir = './ds/test/'\n",
    "\n",
    "classes = [os.path.basename(x[0]) for x in os.walk(train_dir) if x[0]][1:]\n",
    "num_classes = len(classes)\n",
    "print(classes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def load_data(path):\n",
    "    data = []\n",
    "    for c, category in enumerate(classes):\n",
    "        images = [os.path.join(dp, f) for dp, dn, filenames in os.walk(path + category) for f in filenames]\n",
    "\n",
    "        for img_path in images:\n",
    "            img = image.load_img(img_path, target_size=(256, 256))\n",
    "            x = image.img_to_array(img)\n",
    "            x = np.expand_dims(x, axis=0)\n",
    "\n",
    "            data.append({'x' : np.array(x[0]), 'y' : c})\n",
    "\n",
    "    return data\n",
    "\n",
    "\n",
    "data_train_val = load_data(train_dir)\n",
    "data_test = load_data(test_dir)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train images: 872 (872, 21)\n",
      "Val images: 219 (219, 21)\n",
      "Test images: 279 (279, 21)\n"
     ]
    }
   ],
   "source": [
    "X_train_val, y_train_val = np.array([t[\"x\"] for t in data_train_val]), [t[\"y\"] for t in data_train_val]\n",
    "X_test, y_test = np.array([t[\"x\"] for t in data_test]), [t[\"y\"] for t in data_test]\n",
    "\n",
    "X_train, X_val, y_train, y_val = train_test_split(X_train_val, y_train_val, random_state=42, test_size=0.2)\n",
    "\n",
    "# Convert classes to one-hot vectors\n",
    "y_train = to_categorical(y_train)\n",
    "y_val = to_categorical(y_val)\n",
    "y_test = to_categorical(y_test)\n",
    "\n",
    "print('Train images:', len(y_train), y_train.shape)\n",
    "print('Val images:', len(y_val), y_val.shape)\n",
    "print('Test images:', len(y_test), y_test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Enhancement"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Balance classes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The imbalance in the training data is now reduced. Class weights are used to let the CNN pay special attention to classes that are underrepresented."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Compute class weights to let the CNN pay attention to less frequent classes when fitting the model\n",
    "y_integers = np.argmax(y_train, axis=1)\n",
    "class_weights = compute_class_weight(class_weight=\"balanced\",\n",
    "                                     classes=np.unique(y_integers),\n",
    "                                     y=y_integers)\n",
    "class_weights = dict(zip(np.unique(y_integers), class_weights))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Generate artificial images using shift, rotation, flip"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Use an ImageDataGenerator to slightly change images in each epoch. Note that this is only done to training data, but not to validation data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Enrich train data set by rotation, flip, shift\n",
    "gen_train = ImageDataGenerator(rotation_range=5,\n",
    "                                   width_shift_range=0.1,\n",
    "                                   height_shift_range=0.1,\n",
    "                                   horizontal_flip=True,\n",
    "                                   data_format='channels_last',\n",
    "                                   preprocessing_function=preprocess_input)\n",
    "\n",
    "# No augmentation for test and validation sets\n",
    "gen_val_test = ImageDataGenerator(data_format='channels_last',\n",
    "                                  preprocessing_function=preprocess_input)\n",
    "\n",
    "# Generate iterators for train/val/test sets\n",
    "train_iter = gen_train.flow(X_train, y_train, batch_size=16, shuffle=True)\n",
    "val_iter = gen_val_test.flow(X_val, y_val, batch_size=16, shuffle=True)\n",
    "test_iter = gen_val_test.flow(X_test, y_test, batch_size=16, shuffle=True)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model Creation and Training"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Create a CNN model using multiple convolutional layers. As opposed to constant a number of filters in each convolutional layer, this model increases number of filters in deeper layers as more complex features can be learned."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_model():\n",
    "    model = Sequential()\n",
    "    print(\"Input dimensions:\", X_train.shape[1:])\n",
    "\n",
    "    model.add(Conv2D(32, (3, 3), input_shape=X_train.shape[1:]))\n",
    "    model.add(Activation('relu'))\n",
    "    model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "\n",
    "    model.add(Conv2D(32, (3, 3)))\n",
    "    model.add(Activation('relu'))\n",
    "    model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "    model.add(BatchNormalization())\n",
    "\n",
    "    model.add(Dropout(0.2))\n",
    "\n",
    "    model.add(Conv2D(64, (3, 3)))\n",
    "    model.add(Activation('relu'))\n",
    "    model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "    model.add(BatchNormalization())\n",
    "\n",
    "    model.add(Dropout(0.3))\n",
    "\n",
    "    model.add(Conv2D(128, (3, 3)))\n",
    "    model.add(Activation('relu'))\n",
    "    model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "    model.add(BatchNormalization())\n",
    "\n",
    "    model.add(Conv2D(256, (3, 3)))\n",
    "    model.add(Activation('relu'))\n",
    "    model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "    model.add(BatchNormalization())\n",
    "    \n",
    "    model.add(Dropout(0.4))\n",
    "\n",
    "    model.add(Flatten())\n",
    "    model.add(Dense(256))\n",
    "    model.add(Activation('relu'))\n",
    "\n",
    "    model.add(Dropout(0.5))\n",
    "\n",
    "    model.add(Dense(num_classes))\n",
    "    model.add(Activation('softmax'))\n",
    "\n",
    "    # model.summary()\n",
    "\n",
    "    return model\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "baseline_model = get_model()\n",
    "\n",
    "baseline_model.compile(loss='categorical_crossentropy',\n",
    "              optimizer='adam',\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create checkpoints during training\n",
    "filepath = './models/baseline_v2/baseline_v2_{epoch:02d}_{val_accuracy:.2f}.hdf5'\n",
    "\n",
    "checkpoint = ModelCheckpoint(filepath, monitor='val_accuracy', verbose=1, save_best_only=True, mode='max')\n",
    "lr = ReduceLROnPlateau(monitor=\"val_loss\", factor=0.5, patience=3)\n",
    "\n",
    "callbacks_list = [checkpoint, lr]\n",
    "\n",
    "# Train the model\n",
    "history = baseline_model.fit(train_iter,\n",
    "                    # batch_size=32,\n",
    "                    epochs=75,\n",
    "                    validation_data=val_iter,\n",
    "                    callbacks=callbacks_list,\n",
    "                    class_weight=class_weights,\n",
    "                    workers=3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Visualize training progress along epochs"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Plot the (val_)accuracy and (val_)loss along epochs to identify over- and underfitting."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "fig = plt.figure(figsize=(16,4))\n",
    "ax = fig.add_subplot(121)\n",
    "ax.plot(history.history[\"val_loss\"], label='validation')\n",
    "ax.plot(history.history[\"loss\"], label='train')\n",
    "ax.legend(loc=\"upper right\")\n",
    "ax.set_title(\"Loss\")\n",
    "ax.set_xlabel(\"epochs\")\n",
    "\n",
    "ax2 = fig.add_subplot(122)\n",
    "ax2.plot(history.history[\"val_accuracy\"], label='validation')\n",
    "ax2.plot(history.history[\"accuracy\"], label='train')\n",
    "ax2.legend(loc=\"lower right\")\n",
    "ax2.set_title(\"Accuracy\")\n",
    "ax2.set_xlabel(\"epochs\")\n",
    "ax2.set_ylim(0, 1)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Evaluation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Analysis of model performance on test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# NOTE: By default, the finished model is loaded from the repository in order to evaluate\n",
    "# without compiling and fitting the model again. However, if you train again,\n",
    "# uncomment the following line and comment the load_model() call below\n",
    "\n",
    "# test_model = baseline_model\n",
    "test_model = keras.models.load_model('./models/baseline_v2/baseline_v2_epochs=35_acc=47.h5')\n",
    "\n",
    "y_pred = test_model.predict(X_test)\n",
    "y_pred = np.argmax(y_pred, axis = 1)\n",
    "y_true = np.argmax(y_test, axis = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1008x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot the distribution of predicted classes\n",
    "\n",
    "unique, counts = np.unique(y_pred, return_counts=True)\n",
    "ticks = np.arange(len(counts))\n",
    "pred_dist = sorted(zip(classes, counts), key=lambda x: x[1], reverse=True)\n",
    "\n",
    "plt.figure(figsize=(14,6))\n",
    "plt.grid(color='w', axis='y')\n",
    "plt.bar(ticks, [c[1] for c in pred_dist])\n",
    "plt.xticks(ticks, [c[0][:4] for c in pred_dist], rotation=75)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Acc:  0.4731182795698925\n",
      "Prec:  0.44915906180983317\n",
      "Rec:  0.4731182795698925\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 792x792 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot the confusion matrix and evaluate acc, prec and rec\n",
    "\n",
    "cm = confusion_matrix(y_true, y_pred)\n",
    "cm = np.round(cm, 2)\n",
    "\n",
    "cmd = ConfusionMatrixDisplay(cm, display_labels=[c[:4] for c in classes])\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(11,11))\n",
    "cmd.plot(ax=ax, xticks_rotation='vertical')\n",
    "\n",
    "print('Acc: ', accuracy_score(y_true, y_pred))\n",
    "print('Prec: ', precision_score(y_true, y_pred, average='weighted'))\n",
    "print('Rec: ', recall_score(y_true, y_pred, average='weighted'))\n",
    "# Acc:  0.4731182795698925\n",
    "# Prec:  0.44915906180983317\n",
    "# Rec:  0.4731182795698925"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "9975a71c6d7ff9b9547f9907661bca7a033cdc6e7c87f596a27a936a2886f07d"
  },
  "kernelspec": {
   "display_name": "Python 3.8.10 ('venv': venv)",
   "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": "3.8.10"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}