CNN_Scratch_with_balanced_weight.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import Libraries\n",
    "from sklearn.utils import class_weight\n",
    "from keras.preprocessing.image import ImageDataGenerator, load_img\n",
    "import tensorflow as tf\n",
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "import random\n",
    "import os\n",
    "from PIL import Image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "TRAINING_DIR = \"dataset/train_images/\"\n",
    "TEST_DIR = \"dataset/test_images/\"\n",
    "DEFAULT_SIZE = (150, 150)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 2408 images belonging to 10 classes.\n",
      "Found 604 images belonging to 10 classes.\n",
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d_3 (Conv2D)            (None, 148, 148, 16)      448       \n",
      "_________________________________________________________________\n",
      "max_pooling2d_3 (MaxPooling2 (None, 74, 74, 16)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_4 (Conv2D)            (None, 72, 72, 32)        4640      \n",
      "_________________________________________________________________\n",
      "max_pooling2d_4 (MaxPooling2 (None, 36, 36, 32)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_5 (Conv2D)            (None, 34, 34, 64)        18496     \n",
      "_________________________________________________________________\n",
      "max_pooling2d_5 (MaxPooling2 (None, 17, 17, 64)        0         \n",
      "_________________________________________________________________\n",
      "flatten_1 (Flatten)          (None, 18496)             0         \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 512)               9470464   \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              (None, 10)                5130      \n",
      "=================================================================\n",
      "Total params: 9,499,178\n",
      "Trainable params: 9,499,178\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "\n",
    "ImageDataGenerator = tf.keras.preprocessing.image.ImageDataGenerator\n",
    "#Using ImageDataGenerator to ease Data preparation as it lables images based on Folder Name which is ideal for the way Data Set is arranged\n",
    "\n",
    "train_datagen = ImageDataGenerator(rescale=1.0/255.)\n",
    "train_generator = train_datagen.flow_from_directory(TRAINING_DIR,\n",
    "                                                    batch_size=32,\n",
    "                                                    class_mode='binary',\n",
    "                                                    target_size=(150, 150))\n",
    "\n",
    "test_datagen = ImageDataGenerator(rescale=1.0/255.)\n",
    "test_generator = test_datagen.flow_from_directory(TEST_DIR,\n",
    "                                                  batch_size=32,\n",
    "                                                  class_mode='binary',\n",
    "                                                  target_size=(150, 150))\n",
    "model = tf.keras.models.Sequential([\n",
    "    tf.keras.layers.Conv2D(16, (3, 3), activation='relu', input_shape=(150, 150, 3)),\n",
    "    tf.keras.layers.MaxPooling2D(2, 2),\n",
    "    tf.keras.layers.Conv2D(32, (3, 3), activation='relu'),\n",
    "    tf.keras.layers.MaxPooling2D(2, 2),\n",
    "    tf.keras.layers.Conv2D(64, (3, 3), activation='relu'),\n",
    "    tf.keras.layers.MaxPooling2D(2, 2),\n",
    "    tf.keras.layers.Flatten(),\n",
    "    tf.keras.layers.Dense(512, activation='relu'),\n",
    "    tf.keras.layers.Dense(10, activation='softmax')\n",
    "])\n",
    "model.compile(optimizer='adam', loss='sparse_categorical_crossentropy', metrics=['accuracy'])\n",
    "model.summary()\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "class_weight = {\n",
    "        0: (604 / (10 * 160)),\n",
    "        1: (604 / (10 * 26)),\n",
    "        2: (604 / (10 * 19)),\n",
    "        3: (604 / (10 * 25)),\n",
    "        4: (604 / (10 * 87)),\n",
    "        5: (604 / (10 * 23)),\n",
    "        6: (604 / (10 * 40)),\n",
    "        7: (604 / (10 * 160)),\n",
    "        8: (604 / (10 * 17)),\n",
    "        9: (604 / (10 * 47)) \n",
    "    }\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "76/76 [==============================] - 62s 806ms/step - loss: 1.8231 - accuracy: 0.4413 - val_loss: 0.4876 - val_accuracy: 0.8709\n",
      "Epoch 2/10\n",
      "76/76 [==============================] - 61s 800ms/step - loss: 0.5817 - accuracy: 0.8801 - val_loss: 0.1837 - val_accuracy: 0.9503\n",
      "Epoch 3/10\n",
      "76/76 [==============================] - 61s 800ms/step - loss: 0.2285 - accuracy: 0.9555 - val_loss: 0.3004 - val_accuracy: 0.9139\n",
      "Epoch 4/10\n",
      "76/76 [==============================] - 65s 852ms/step - loss: 0.2102 - accuracy: 0.9630 - val_loss: 0.2205 - val_accuracy: 0.9387\n",
      "Epoch 5/10\n",
      "76/76 [==============================] - 68s 899ms/step - loss: 0.0674 - accuracy: 0.9876 - val_loss: 0.2911 - val_accuracy: 0.9238\n",
      "Epoch 6/10\n",
      "76/76 [==============================] - 70s 920ms/step - loss: 0.0512 - accuracy: 0.9849 - val_loss: 0.2560 - val_accuracy: 0.9404\n",
      "Epoch 7/10\n",
      "76/76 [==============================] - 73s 964ms/step - loss: 0.0415 - accuracy: 0.9869 - val_loss: 0.2328 - val_accuracy: 0.9470\n",
      "Epoch 8/10\n",
      "76/76 [==============================] - 64s 842ms/step - loss: 0.0092 - accuracy: 0.9982 - val_loss: 0.2262 - val_accuracy: 0.9487\n",
      "Epoch 9/10\n",
      "76/76 [==============================] - 62s 819ms/step - loss: 0.0019 - accuracy: 1.0000 - val_loss: 0.2199 - val_accuracy: 0.9570\n",
      "Epoch 10/10\n",
      "76/76 [==============================] - 65s 853ms/step - loss: 6.1534e-04 - accuracy: 1.0000 - val_loss: 0.2154 - val_accuracy: 0.9603\n"
     ]
    }
   ],
   "source": [
    "history = model.fit(train_generator,\n",
    "                    epochs=10,\n",
    "                    verbose=1,\n",
    "                    validation_data=test_generator,class_weight= class_weight)\n",
    "                              \n",
    "model.save('CNNmodel.h5')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "execution_count": 26,
     "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"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Visualisation\n",
    "import matplotlib.image  as mpimg\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "#-----------------------------------------------------------\n",
    "# Retrieve a list of list results on training and test data\n",
    "# sets for each training epoch\n",
    "#-----------------------------------------------------------\n",
    "acc=history.history['accuracy']\n",
    "val_acc=history.history['val_accuracy']\n",
    "loss=history.history['loss']\n",
    "val_loss=history.history['val_loss']\n",
    "\n",
    "epochs=range(1,11) # Get number of epochs\n",
    "#------------------------------------------------\n",
    "# Plot training and validation accuracy per epoch\n",
    "#------------------------------------------------\n",
    "plt.plot(epochs, acc, 'r')\n",
    "plt.plot(epochs, val_acc, 'b')\n",
    "plt.title('Training and Testing accuracy')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['train', 'test'], loc='lower right')\n",
    "plt.figure()\n",
    "#------------------------------------------------\n",
    "# Plot training and validation loss per epoch\n",
    "#------------------------------------------------\n",
    "plt.plot(epochs, loss, 'r')\n",
    "plt.plot(epochs, val_loss, 'b')\n",
    "plt.legend(['train', 'test'], loc='upper right')\n",
    "plt.title('Training and Testing Loss')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.figure()\n",
    "# Desired output. Charts with training and validation metrics. No crash :)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "19/19 [==============================] - 5s 249ms/step - loss: 0.2154 - accuracy: 0.9603\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.21542823314666748, 0.9602649211883545]"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.evaluate(test_generator)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['Airplanes', 'Bonsai', 'Butterfly', 'Car_side', 'Faces', 'ketch', 'Leopards', 'Motorbikes', 'Scorpion', 'Watch']\n"
     ]
    }
   ],
   "source": [
    "classes = [\"Airplanes\",\"Bonsai\",\"Butterfly\",\"Car_side\",\"Faces\",\"ketch\",\"Leopards\",\"Motorbikes\",\"Scorpion\",\"Watch\"]\n",
    "true_classes = test_generator.classes\n",
    "print(classes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Class 0: 26.359%, Class 1: 4.283%, Class 2: 3.130%, Class 3: 4.119%, Class 4: 14.333%,Class 5: 3.789%, Class 6: 7.084%, Class 7: 26.359%, Class 8: 2.801%,Class 9: 7.743%\n"
     ]
    },
    {
     "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": [
    "# evaluate a majority class classifier on an 1:100 imbalanced dataset\n",
    "from numpy import mean\n",
    "from numpy import hstack\n",
    "from numpy import vstack\n",
    "from numpy import where\n",
    "from matplotlib import pyplot\n",
    "from sklearn.datasets import make_blobs\n",
    "from sklearn.dummy import DummyClassifier\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.model_selection import RepeatedStratifiedKFold\n",
    " \n",
    "# create a dataset with a given class distribution\n",
    "def get_dataset(proportions):\n",
    "\t# determine the number of classes\n",
    "\tn_classes = 10\n",
    "\t# determine the number of examples to generate for each class\n",
    "\tlargest = max([v for k,v in proportions.items()])\n",
    "\tn_samples = largest * n_classes\n",
    "\t# create dataset\n",
    "\tX, y = make_blobs(n_samples=n_samples, centers=n_classes, n_features=2, random_state=1, cluster_std=3)\n",
    "\t# collect the examples\n",
    "\tX_list, y_list = list(), list()\n",
    "\tfor k,v in proportions.items():\n",
    "\t\trow_ix = where(y == k)[0]\n",
    "\t\tselected = row_ix[:v]\n",
    "\t\tX_list.append(X[selected, :])\n",
    "\t\ty_list.append(y[selected])\n",
    "\treturn vstack(X_list), hstack(y_list)\n",
    " \n",
    "# evaluate a model using repeated k-fold cross-validation\n",
    "def evaluate_model(X, y, metric):\n",
    "\t# define model\n",
    "\tmodel = DummyClassifier(strategy='most_frequent')\n",
    "\t# evaluate a model with repeated stratified k fold cv\n",
    "\tcv = RepeatedStratifiedKFold(n_splits=10, n_repeats=3, random_state=1)\n",
    "\tscores = cross_val_score(model, X, y, scoring=metric, cv=cv, n_jobs=-1)\n",
    "\treturn scores\n",
    "# scatter plot of dataset, different color for each class\n",
    "def plot_dataset(X, y):\n",
    "\t# create scatter plot for samples from each class\n",
    "\tn_classes = 10\n",
    "\tfor class_value in range(n_classes):\n",
    "\t\t# get row indexes for samples with this class\n",
    "\t\trow_ix = where(y == class_value)[0]\n",
    "\t\t# create scatter of these samples\n",
    "\t\tpyplot.scatter(X[row_ix, 0], X[row_ix, 1], label=str(class_value))\n",
    "\t# show a legend\n",
    "\tpyplot.legend()\n",
    "\t# show the plot\n",
    "\tpyplot.show()\n",
    "# define the class distribution 1:100\n",
    "proportions = {0:160, 1:26,2:19, 3:25,4:87, 5:23,6:43, 7:160,8:17, 9:47}\n",
    "# generate dataset\n",
    "X, y = get_dataset(proportions)\n",
    "# summarize class distribution:\n",
    "major1 = (len(where(y == 0)[0]) / len(X)) * 100\n",
    "minor1 = (len(where(y == 1)[0]) / len(X)) * 100\n",
    "major2 = (len(where(y == 2)[0]) / len(X)) * 100\n",
    "minor2 = (len(where(y == 3)[0]) / len(X)) * 100\n",
    "major3 = (len(where(y == 4)[0]) / len(X)) * 100\n",
    "minor3 = (len(where(y == 5)[0]) / len(X)) * 100\n",
    "major4 = (len(where(y == 6)[0]) / len(X)) * 100\n",
    "minor4 = (len(where(y == 7)[0]) / len(X)) * 100\n",
    "major5 = (len(where(y == 8)[0]) / len(X)) * 100\n",
    "minor5 = (len(where(y == 9)[0]) / len(X)) * 100\n",
    "print('Class 0: %.3f%%, Class 1: %.3f%%, Class 2: %.3f%%, Class 3: %.3f%%, Class 4: %.3f%%,Class 5: %.3f%%, Class 6: %.3f%%, Class 7: %.3f%%, Class 8: %.3f%%,Class 9: %.3f%%'\n",
    "      %(major1, minor1,major2, minor2,major3, minor3,major4, minor4,major5, minor5))\n",
    "plot_dataset(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'airplanes': 0,\n",
       " 'bonsai': 1,\n",
       " 'butterfly': 2,\n",
       " 'car_side': 3,\n",
       " 'faces': 4,\n",
       " 'ketch': 5,\n",
       " 'leopards': 6,\n",
       " 'motorbikes': 7,\n",
       " 'scorpion': 8,\n",
       " 'watch': 9}"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "test_generator.class_indices"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.preprocessing import image as image_utils\n",
    "image = keras.preprocessing.image\n",
    "path = 'dataset/test_images/faces/image_0389.jpg'\n",
    "img = image.load_img(path, target_size=(150, 150))\n",
    "test_image = image_utils.img_to_array(img)\n",
    "test_image = np.expand_dims(test_image, axis=0)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "4\n",
      "faces\n"
     ]
    }
   ],
   "source": [
    "result = model.predict_on_batch(test_image)\n",
    "# Get the filenames from the generator\n",
    "fnames = test_generator.filenames\n",
    "\n",
    "ground_truth = test_generator.classes\n",
    "label2index = test_generator.class_indices\n",
    "idx2label = dict((v,k) for k,v in label2index.items())\n",
    "print(result.argmax())\n",
    "for category, value in test_generator.class_indices.items():\n",
    "            if value == result.argmax():\n",
    "                print(category)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 604 images belonging to 10 classes.\n",
      "7/7 [==============================] - 5s 623ms/step\n",
      "No of errors = 24/604\n"
     ]
    }
   ],
   "source": [
    "# Create a generator for prediction\n",
    "tag_size =(150,150)\n",
    "test_generator = test_datagen.flow_from_directory(\n",
    "        TEST_DIR,\n",
    "        target_size=tag_size,\n",
    "        batch_size=100,\n",
    "        class_mode='categorical',\n",
    "        shuffle=False)\n",
    "# Get the filenames from the generator\n",
    "fnames = test_generator.filenames\n",
    "\n",
    "ground_truth = test_generator.classes\n",
    "label2index = test_generator.class_indices\n",
    "idx2label = dict((v,k) for k,v in label2index.items())\n",
    "       \n",
    "# Get the predictions from the model using the generator\n",
    "predictions = model.predict(test_generator,verbose=1)\n",
    "predicted_classes = np.argmax(predictions,axis=1)\n",
    "\n",
    "errors = np.where(predicted_classes != ground_truth)[0]\n",
    "print(\"No of errors = {}/{}\".format(len(errors),test_generator.samples))\n",
    "\n",
    "# Show the errors\n",
    "# for i in range(len(errors)):\n",
    "#     pred_class = np.argmax(predictions[errors[i]])\n",
    "#     print(\"Predicted class:\",pred_class)\n",
    "#     pred_label = idx2label[pred_class]\n",
    "#     print(\"Predicted label:\",pred_label)\n",
    "#     title = 'Original label:{}, Prediction :{}, confidence : {:.3f}'.format(\n",
    "#         fnames[errors[i]].split('/')[0],\n",
    "#         pred_label,\n",
    "#         predictions[errors[i]][pred_class])\n",
    "    \n",
    "#     original = load_img('{}/{}'.format(TEST_DIR,fnames[errors[i]]))\n",
    "#     plt.figure(figsize=[7,7])\n",
    "#     plt.axis('off')\n",
    "#     plt.title(title)\n",
    "#     plt.imshow(original)\n",
    "#     plt.show()\n",
    "#     label2index = test_generator.class_indices\n",
    "# # Getting the mapping from class index to class label\n",
    "# idx2label = dict((v,k) for k,v in label2index.items())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "def PlotConfusionMatrix(ground_truth, predicted_classes):\n",
    "    np.set_printoptions(precision=2)\n",
    "    classes = ['airplanes','bonsai','butterfly','car_side','faces','ketch','leopards','motorbikes','scorpion','watch']\n",
    "    title='Confusion matrix'\n",
    "    cmap=plt.cm.Blues\n",
    "    cm = confusion_matrix(ground_truth,predicted_classes)\n",
    "    fig, ax = plt.subplots()\n",
    "    im = ax.imshow(cm, interpolation='nearest', cmap=cmap)\n",
    "    ax.figure.colorbar(im, ax=ax)\n",
    "    ax.set(xticks=np.arange(cm.shape[1]),\n",
    "           yticks=np.arange(cm.shape[0]),\n",
    "           xticklabels=classes, yticklabels=classes,\n",
    "           title=title,\n",
    "           ylabel='True label',\n",
    "           xlabel='Predicted label')\n",
    "    plt.setp(ax.get_xticklabels(), rotation=45, ha=\"right\",rotation_mode=\"anchor\")\n",
    "    fmt ='d' \n",
    "    thresh = cm.max() / 2.\n",
    "    for i in range(cm.shape[0]):\n",
    "        for j in range(cm.shape[1]):\n",
    "            ax.text(j, i, format(cm[i, j],fmt),\n",
    "                    ha=\"center\", va=\"center\",\n",
    "                    color=\"white\" if cm[i, j] > thresh else \"black\")\n",
    "    fig.tight_layout()    \n",
    "    plt.show()\n",
    "PlotConfusionMatrix(ground_truth, predicted_classes)"
   ]
  },
  {
   "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": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}