model_training.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import cv2\n",
    "from tensorflow import keras\n",
    "import tensorflow as tf\n",
    "import os"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.utils import shuffle\n",
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow.keras.utils as utils"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense , Activation , Dropout ,Flatten,  MaxPooling2D, Conv2D"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    " data_path = 'data/ckplus/CK+48'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "7 classes are ['anger', 'contempt', 'disgust', 'fear', 'happy', 'sadness', 'surprise']\n"
     ]
    }
   ],
   "source": [
    "data_dir_list = os.listdir(data_path)\n",
    "#data_dir_list\n",
    "print(str(len(data_dir_list))+' classes are',data_dir_list)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "number of images in anger =  135\n",
      "number of images in contempt =  54\n",
      "number of images in disgust =  177\n",
      "number of images in fear =  75\n",
      "number of images in happy =  207\n",
      "number of images in sadness =  84\n",
      "number of images in surprise =  249\n"
     ]
    }
   ],
   "source": [
    "img_data = []\n",
    "for dataset in data_dir_list:\n",
    "    img_list = os.listdir(data_path+'/'+dataset)\n",
    "    print('number of images in '+str(dataset)+' = ',len(img_list))\n",
    "    for img in img_list:\n",
    "        img = cv2.imread(data_path+'/'+dataset+'/'+img)\n",
    "        \n",
    "        img = cv2.resize(img,(48,48))\n",
    "        \n",
    "        img_data.append(img)\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(48, 48, 3)"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "img_data = np.array(img_data)\n",
    "img_data[0].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(981, 48, 48, 3)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "img_data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Normalization (essential for neural nets)\n",
    "img_data = img_data.astype('float32')\n",
    "img_data = img_data/255"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(981, 48, 48, 3)"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "img_data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "shape of image (48, 48)\n",
      "num of images  981\n"
     ]
    }
   ],
   "source": [
    "print('shape of image', (img_data.shape[1],img_data.shape[2]))\n",
    "print('num of images ', img_data.shape[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "label_text = {0:'anger', 1:'contempt', 2: 'contempt',\n",
    "                         3:'fear', 4:'happiness',\n",
    "                         5: 'sadness', 6: 'surprise'}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "labels = np.ones((img_data.shape[0],),dtype='int64')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "labels[0:134]=0 #135\n",
    "labels[135:188]=1 #54\n",
    "labels[189:365]=2 #177\n",
    "\n",
    "labels[366:440]=3 #75\n",
    "labels[441:647]=4 #207\n",
    "\n",
    "labels[648:731]=5 #84\n",
    "labels[732:980]=6 #249"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plotImages(images_arr):\n",
    "    fig,axes = plt.subplots(1,1,figsize=(5,5))\n",
    "#     axes = axes.flatten()\n",
    "    axes.imshow(images_arr)\n",
    "    axes.axis('off')\n",
    "    plt.tight_layout()\n",
    "    plt.show()\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "happiness\n"
     ]
    }
   ],
   "source": [
    "plotImages(img_data[600])\n",
    "print(label_text[labels[600]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "def create_model(optim):\n",
    "    input_shape=(48,48,3)\n",
    "\n",
    "    model = Sequential()\n",
    "    model.add(Conv2D(6, (5, 5), input_shape=input_shape, padding='same', activation = 'relu'))\n",
    "    model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "\n",
    "    model.add(Conv2D(16, (5, 5), padding='same', activation = 'relu'))\n",
    "    model.add(Activation('relu'))\n",
    "    model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "\n",
    "    model.add(Conv2D(64, (3, 3), activation = 'relu'))\n",
    "    model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "\n",
    "    model.add(Flatten())\n",
    "    model.add(Dense(128, activation = 'relu'))\n",
    "    model.add(Dropout(0.5))\n",
    "    model.add(Dense(7, activation = 'softmax'))\n",
    "\n",
    "    model.compile(loss='categorical_crossentropy', metrics=['accuracy'],optimizer=optim)\n",
    "    \n",
    "    return model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
    "from tensorflow.keras.callbacks import ModelCheckpoint, EarlyStopping, ReduceLROnPlateau\n",
    "from sklearn.model_selection import cross_val_predict, cross_val_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "aug = ImageDataGenerator(rotation_range=25, width_shift_range=0.1,\n",
    "                        height_shift_range=0.1, shear_range=0.2,\n",
    "                        zoom_range=0.2, horizontal_flip=True,\n",
    "                        fill_mode='nearrest')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.optimizers import *"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "x,y = shuffle(img_data,Y,random_state = 3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(x, y, test_size=0.2, random_state=2)\n",
    "x_test=X_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "981\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([[1., 0., 0., ..., 0., 0., 0.],\n",
       "       [1., 0., 0., ..., 0., 0., 0.],\n",
       "       [1., 0., 0., ..., 0., 0., 0.],\n",
       "       ...,\n",
       "       [0., 0., 0., ..., 0., 0., 1.],\n",
       "       [0., 0., 0., ..., 0., 0., 1.],\n",
       "       [0., 1., 0., ..., 0., 0., 0.]], dtype=float32)"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Y = utils.to_categorical(labels, len(dataset)-1)\n",
    "print(len(Y))\n",
    "Y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "981"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(img_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0. 0. 0. 0. 0. 0. 1.]\n"
     ]
    }
   ],
   "source": [
    "plotImages(x[0])\n",
    "print(y[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "# from sklearn.model_selection import KFold"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "# kf = KFold(n_splits=5, shuffle=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "aug.fit(X_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "early_stopping = EarlyStopping(\n",
    "    monitor='val_accuracy',\n",
    "    min_delta=0.00005,\n",
    "    patience=11,\n",
    "    verbose=1,\n",
    "    restore_best_weights=True,\n",
    ")\n",
    "\n",
    "lr_scheduler = ReduceLROnPlateau(\n",
    "    monitor='val_accuracy',\n",
    "    factor=0.5,\n",
    "    patience=7,\n",
    "    min_lr=1e-7,\n",
    "    verbose=1,\n",
    ")\n",
    "\n",
    "callbacks = [\n",
    "    early_stopping,\n",
    "    lr_scheduler,\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "batch_size = 32 #batch size of 32 performs the best.\n",
    "epochs = 100\n",
    "optims = [\n",
    "    Nadam(learning_rate=0.001, beta_1=0.9, beta_2=0.999, epsilon=1e-07, name='Nadam'),\n",
    "    Adam(0.001),\n",
    "]\n",
    "\n",
    "\n",
    "model = create_model(optims[1]) \n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d (Conv2D)              (None, 48, 48, 6)         456       \n",
      "_________________________________________________________________\n",
      "max_pooling2d (MaxPooling2D) (None, 24, 24, 6)         0         \n",
      "_________________________________________________________________\n",
      "conv2d_1 (Conv2D)            (None, 24, 24, 16)        2416      \n",
      "_________________________________________________________________\n",
      "activation (Activation)      (None, 24, 24, 16)        0         \n",
      "_________________________________________________________________\n",
      "max_pooling2d_1 (MaxPooling2 (None, 12, 12, 16)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_2 (Conv2D)            (None, 10, 10, 64)        9280      \n",
      "_________________________________________________________________\n",
      "max_pooling2d_2 (MaxPooling2 (None, 5, 5, 64)          0         \n",
      "_________________________________________________________________\n",
      "flatten (Flatten)            (None, 1600)              0         \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 128)               204928    \n",
      "_________________________________________________________________\n",
      "dropout (Dropout)            (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 7)                 903       \n",
      "=================================================================\n",
      "Total params: 217,983\n",
      "Trainable params: 217,983\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n",
      "23/23 - 1s - loss: 1.8314 - accuracy: 0.2567 - val_loss: 1.7486 - val_accuracy: 0.2532\n",
      "Epoch 2/100\n",
      "23/23 - 1s - loss: 1.7708 - accuracy: 0.3121 - val_loss: 1.6767 - val_accuracy: 0.3165\n",
      "Epoch 3/100\n",
      "23/23 - 1s - loss: 1.6179 - accuracy: 0.4043 - val_loss: 1.4187 - val_accuracy: 0.4557\n",
      "Epoch 4/100\n",
      "23/23 - 1s - loss: 1.3359 - accuracy: 0.5092 - val_loss: 1.1353 - val_accuracy: 0.5823\n",
      "Epoch 5/100\n",
      "23/23 - 1s - loss: 1.1482 - accuracy: 0.5589 - val_loss: 0.9913 - val_accuracy: 0.6329\n",
      "Epoch 6/100\n",
      "23/23 - 1s - loss: 0.9731 - accuracy: 0.6369 - val_loss: 0.8014 - val_accuracy: 0.7722\n",
      "Epoch 7/100\n",
      "23/23 - 1s - loss: 0.8787 - accuracy: 0.6723 - val_loss: 0.7090 - val_accuracy: 0.7848\n",
      "Epoch 8/100\n",
      "23/23 - 1s - loss: 0.8417 - accuracy: 0.6950 - val_loss: 0.6416 - val_accuracy: 0.7342\n",
      "Epoch 9/100\n",
      "23/23 - 1s - loss: 0.6714 - accuracy: 0.7433 - val_loss: 0.5970 - val_accuracy: 0.8228\n",
      "Epoch 10/100\n",
      "23/23 - 1s - loss: 0.6323 - accuracy: 0.7702 - val_loss: 0.5719 - val_accuracy: 0.8861\n",
      "Epoch 11/100\n",
      "23/23 - 1s - loss: 0.5814 - accuracy: 0.8099 - val_loss: 0.5362 - val_accuracy: 0.8608\n",
      "Epoch 12/100\n",
      "23/23 - 1s - loss: 0.5713 - accuracy: 0.8028 - val_loss: 0.4723 - val_accuracy: 0.9241\n",
      "Epoch 13/100\n",
      "23/23 - 1s - loss: 0.4466 - accuracy: 0.8610 - val_loss: 0.4912 - val_accuracy: 0.8987\n",
      "Epoch 14/100\n",
      "23/23 - 1s - loss: 0.4108 - accuracy: 0.8624 - val_loss: 0.3804 - val_accuracy: 0.9494\n",
      "Epoch 15/100\n",
      "23/23 - 1s - loss: 0.3690 - accuracy: 0.8624 - val_loss: 0.4051 - val_accuracy: 0.9367\n",
      "Epoch 16/100\n",
      "23/23 - 1s - loss: 0.3296 - accuracy: 0.8879 - val_loss: 0.3835 - val_accuracy: 0.9367\n",
      "Epoch 17/100\n",
      "23/23 - 1s - loss: 0.3513 - accuracy: 0.8837 - val_loss: 0.3839 - val_accuracy: 0.9241\n",
      "Epoch 18/100\n",
      "23/23 - 1s - loss: 0.2756 - accuracy: 0.9135 - val_loss: 0.4362 - val_accuracy: 0.9241\n",
      "Epoch 19/100\n",
      "23/23 - 1s - loss: 0.2753 - accuracy: 0.9191 - val_loss: 0.3398 - val_accuracy: 0.9367\n",
      "Epoch 20/100\n",
      "23/23 - 1s - loss: 0.4643 - accuracy: 0.8184 - val_loss: 0.3630 - val_accuracy: 0.9747\n",
      "Epoch 21/100\n",
      "23/23 - 1s - loss: 0.3619 - accuracy: 0.8922 - val_loss: 0.3340 - val_accuracy: 0.9367\n",
      "Epoch 22/100\n",
      "23/23 - 1s - loss: 0.2598 - accuracy: 0.9163 - val_loss: 0.3423 - val_accuracy: 0.9747\n",
      "Epoch 23/100\n",
      "23/23 - 1s - loss: 0.2132 - accuracy: 0.9319 - val_loss: 0.3287 - val_accuracy: 0.9747\n",
      "Epoch 24/100\n",
      "23/23 - 1s - loss: 0.2056 - accuracy: 0.9333 - val_loss: 0.3572 - val_accuracy: 0.9494\n",
      "Epoch 25/100\n",
      "23/23 - 1s - loss: 0.2133 - accuracy: 0.9433 - val_loss: 0.3186 - val_accuracy: 0.9747\n",
      "Epoch 26/100\n",
      "23/23 - 1s - loss: 0.1757 - accuracy: 0.9546 - val_loss: 0.3395 - val_accuracy: 0.9747\n",
      "Epoch 27/100\n",
      "\n",
      "Epoch 00027: ReduceLROnPlateau reducing learning rate to 0.0005000000237487257.\n",
      "23/23 - 1s - loss: 0.1662 - accuracy: 0.9532 - val_loss: 0.3020 - val_accuracy: 0.9747\n",
      "Epoch 28/100\n",
      "23/23 - 1s - loss: 0.1591 - accuracy: 0.9589 - val_loss: 0.3420 - val_accuracy: 0.9747\n",
      "Epoch 29/100\n",
      "23/23 - 1s - loss: 0.1355 - accuracy: 0.9603 - val_loss: 0.3498 - val_accuracy: 0.9747\n",
      "Epoch 30/100\n",
      "23/23 - 1s - loss: 0.1412 - accuracy: 0.9560 - val_loss: 0.3421 - val_accuracy: 0.9747\n",
      "Epoch 31/100\n",
      "Restoring model weights from the end of the best epoch.\n",
      "23/23 - 1s - loss: 0.1216 - accuracy: 0.9702 - val_loss: 0.3486 - val_accuracy: 0.9747\n",
      "Epoch 00031: early stopping\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x14fc93be5f8>"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(x=X_train, y= y_train, validation_split=0.1, batch_size=32, \n",
    "          epochs=100, shuffle=True, verbose=2, callbacks=callbacks \n",
    "         )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From <ipython-input-39-25e6cc6e3d40>:2: Sequential.predict_classes (from tensorflow.python.keras.engine.sequential) is deprecated and will be removed after 2021-01-01.\n",
      "Instructions for updating:\n",
      "Please use instead:* `np.argmax(model.predict(x), axis=-1)`,   if your model does multi-class classification   (e.g. if it uses a `softmax` last-layer activation).* `(model.predict(x) > 0.5).astype(\"int32\")`,   if your model does binary classification   (e.g. if it uses a `sigmoid` last-layer activation).\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "results = model.predict_classes(X_test)\n",
    "cm = confusion_matrix(np.where(y_test == 1)[1], results)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "import seaborn as sns\n",
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "label_mapdisgust = ['anger','contempt','disgust','fear','happy','sadness','surprise']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "cm_df = pd.DataFrame(cm, index = label_mapdisgust,\n",
    "                     columns = label_mapdisgust\n",
    "                    )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize = (5,5))\n",
    "sns.heatmap(cm, annot = True,cmap='Greys',cbar=False,linewidth=2,fmt='d')\n",
    "plt.title('CNN Emotion Classify')\n",
    "plt.ylabel('True class')\n",
    "plt.xlabel('Prediction class')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test Loss: 0.30213403701782227\n",
      "Test accuracy: 0.8984771370887756\n",
      "(1, 48, 48, 3)\n",
      "[[8.39535713e-01 7.76557578e-03 1.24645315e-01 1.83928394e-04\n",
      "  2.53597368e-03 2.52125170e-02 1.21115154e-04]]\n",
      "[0]\n",
      "[[1. 0. 0. 0. 0. 0. 0.]]\n"
     ]
    }
   ],
   "source": [
    "score = model.evaluate(X_test, y_test, verbose=0)\n",
    "print('Test Loss:', score[0])\n",
    "print('Test accuracy:', score[1])\n",
    "\n",
    "test_image = X_test[3:4]\n",
    "print (test_image.shape)\n",
    "\n",
    "print(model.predict(test_image))\n",
    "print(model.predict_classes(test_image))\n",
    "print(y_test[3:4])\n",
    "\n",
    "#predict\n",
    "y_pred = model.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    " from sklearn.metrics import multilabel_confusion_matrix\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[[168,   6],\n",
       "        [  3,  20]],\n",
       "\n",
       "       [[188,   0],\n",
       "        [  6,   3]],\n",
       "\n",
       "       [[165,   1],\n",
       "        [  0,  31]],\n",
       "\n",
       "       [[180,   0],\n",
       "        [  8,   9]],\n",
       "\n",
       "       [[158,   1],\n",
       "        [  0,  38]],\n",
       "\n",
       "       [[176,   0],\n",
       "        [ 11,  10]],\n",
       "\n",
       "       [[139,   0],\n",
       "        [  1,  57]]], dtype=int64)"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred = (y_pred>0.5)\n",
    "multilabel_confusion_matrix(y_test, y_pred )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "number of images in anger =  135\n",
      "number of images in contempt =  54\n",
      "number of images in disgust =  177\n",
      "number of images in fear =  75\n",
      "number of images in happy =  207\n",
      "number of images in sadness =  84\n",
      "number of images in surprise =  249\n"
     ]
    }
   ],
   "source": [
    "for dataset in data_dir_list:\n",
    "    img_list = os.listdir(data_path+'/'+dataset)\n",
    "    print('number of images in '+str(dataset)+' = ',len(img_list))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "anger\n"
     ]
    }
   ],
   "source": [
    "plotImages(X_test[3])\n",
    "ans = int(model.predict_classes(test_image)[0])\n",
    "print(label_text[ans])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "if os.path.isfile('model.h5') is False:\n",
    "    model.save('model.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1, 48, 48, 3)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "test_img = X_test[1:2]\n",
    "print(test_img.shape)\n",
    "plotImages(X_test[1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([5], dtype=int64)"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.predict_classes(test_img)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0., 0., 0., 0., 0., 1., 0.]], dtype=float32)"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_test[1:2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[]"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xx = []\n",
    "xx"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "import imageio"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "pic = imageio.imread('ex7.jpeg')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(184, 273, 3)"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pic.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "pic = cv2.resize(pic,(48,48))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "xx.append(pic)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "xx = np.array(xx)\n",
    "xx = xx.astype('float32')\n",
    "xx = xx/255"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1, 48, 48, 3)"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xx.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.predict_classes(xx)[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plotImages(pic)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['anger', 'contempt', 'disgust', 'fear', 'happy', 'sadness', 'surprise']"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_dir_list"
   ]
  }
 ],
 "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": "3.7.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}