{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
    "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5"
   },
   "outputs": [],
   "source": [
    "\n",
    "import numpy as np \n",
    "import pandas as pd \n",
    "import cv2\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "_cell_guid": "79c7e3d0-c299-4dcb-8224-4455121ee9b0",
    "_uuid": "d629ff2d2480ee46fbb7e2d37f6b5fab8052498a"
   },
   "outputs": [],
   "source": [
    "df = pd.read_csv('../input/facial-expression/fer2013.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "len(df.iloc[0]['pixels'].split())\n",
    "# 48 * 48"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "label_map = ['Anger', 'Neutral', 'Fear', 'Happy', 'Sad', 'Surprise']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "img = df.iloc[0]['pixels'].split()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "img = [int(i) for i in img]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "type(img[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "len(img)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "img = np.array(img)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "img = img.reshape(48,48)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "img.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "plt.imshow(img, cmap='gray')\n",
    "plt.xlabel(df.iloc[0]['emotion'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = []\n",
    "y = []"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def getData(path):\n",
    "    anger = 0\n",
    "    fear = 0\n",
    "    sad = 0\n",
    "    happy = 0\n",
    "    surprise = 0\n",
    "    neutral = 0\n",
    "    df = pd.read_csv(path)\n",
    "    \n",
    "    X = []\n",
    "    y = []    \n",
    "    \n",
    "    for i in range(len(df)):\n",
    "        if df.iloc[i]['emotion'] != 1:\n",
    "            if df.iloc[i]['emotion'] == 0:\n",
    "                if anger <= 4000:            \n",
    "                    y.append(df.iloc[i]['emotion'])\n",
    "                    im = df.iloc[i]['pixels']\n",
    "                    im = [int(x) for x in im.split()]\n",
    "                    X.append(im)\n",
    "                    anger += 1\n",
    "                else:\n",
    "                    pass\n",
    "                \n",
    "            if df.iloc[i]['emotion'] == 2:\n",
    "                if fear <= 4000:            \n",
    "                    y.append(df.iloc[i]['emotion'])\n",
    "                    im = df.iloc[i]['pixels']\n",
    "                    im = [int(x) for x in im.split()]\n",
    "                    X.append(im)\n",
    "                    fear += 1\n",
    "                else:\n",
    "                    pass\n",
    "                \n",
    "            if df.iloc[i]['emotion'] == 3:\n",
    "                if happy <= 4000:            \n",
    "                    y.append(df.iloc[i]['emotion'])\n",
    "                    im = df.iloc[i]['pixels']\n",
    "                    im = [int(x) for x in im.split()]\n",
    "                    X.append(im)\n",
    "                    happy += 1\n",
    "                else:\n",
    "                    pass\n",
    "                \n",
    "            if df.iloc[i]['emotion'] == 4:\n",
    "                if sad <= 4000:            \n",
    "                    y.append(df.iloc[i]['emotion'])\n",
    "                    im = df.iloc[i]['pixels']\n",
    "                    im = [int(x) for x in im.split()]\n",
    "                    X.append(im)\n",
    "                    sad += 1\n",
    "                else:\n",
    "                    pass\n",
    "                \n",
    "            if df.iloc[i]['emotion'] == 5:\n",
    "                if surprise <= 4000:            \n",
    "                    y.append(df.iloc[i]['emotion'])\n",
    "                    im = df.iloc[i]['pixels']\n",
    "                    im = [int(x) for x in im.split()]\n",
    "                    X.append(im)\n",
    "                    surprise += 1\n",
    "                else:\n",
    "                    pass\n",
    "                \n",
    "            if df.iloc[i]['emotion'] == 6:\n",
    "                if neutral <= 4000:            \n",
    "                    y.append(df.iloc[i]['emotion'])\n",
    "                    im = df.iloc[i]['pixels']\n",
    "                    im = [int(x) for x in im.split()]\n",
    "                    X.append(im)\n",
    "                    neutral += 1\n",
    "                else:\n",
    "                    pass\n",
    "\n",
    "            \n",
    "            \n",
    "    return X, y  \n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "X, y = getData('../input/facial-expression/fer2013.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.unique(y, return_counts=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = np.array(X)/255.0\n",
    "y = np.array(y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "X.shape, y.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_o = []\n",
    "for i in y:\n",
    "    if i != 6:\n",
    "        y_o.append(i)\n",
    "        \n",
    "    else:\n",
    "        y_o.append(1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.unique(y_o, return_counts=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "for i in range(5):\n",
    "    r = np.random.randint((1), 24000, 1)[0]\n",
    "    plt.figure()\n",
    "    plt.imshow(X[r].reshape(48,48), cmap='gray')\n",
    "    plt.xlabel(label_map[y_o[r]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = X.reshape(len(X), 48, 48, 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# no_of_images, height, width, coloar_map"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "X.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.utils import to_categorical\n",
    "y_new = to_categorical(y_o, num_classes=6)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "len(y_o), y_new.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_o[150], y_new[150]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Dense , Activation , Dropout ,Flatten\n",
    "from keras.layers.convolutional import Conv2D\n",
    "from keras.layers.convolutional import MaxPooling2D\n",
    "from keras.metrics import categorical_accuracy\n",
    "from keras.models import model_from_json\n",
    "from keras.callbacks import ModelCheckpoint\n",
    "from keras.optimizers import *\n",
    "from keras.layers.normalization import BatchNormalization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = Sequential()\n",
    "\n",
    "\n",
    "input_shape = (48,48,1)\n",
    "\n",
    "\n",
    "model.add(Conv2D(64, (5, 5), input_shape=input_shape,activation='relu', padding='same'))\n",
    "model.add(Conv2D(64, (5, 5), padding='same'))\n",
    "model.add(BatchNormalization())\n",
    "model.add(Activation('relu'))\n",
    "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "\n",
    "\n",
    "model.add(Conv2D(128, (5, 5),activation='relu',padding='same'))\n",
    "model.add(Conv2D(128, (5, 5),padding='same'))\n",
    "model.add(BatchNormalization())\n",
    "model.add(Activation('relu'))\n",
    "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "\n",
    "model.add(Conv2D(256, (3, 3),activation='relu',padding='same'))\n",
    "model.add(Conv2D(256, (3, 3),activation='relu',padding='same'))\n",
    "model.add(BatchNormalization())\n",
    "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "\n",
    "## (15, 15) --->  30\n",
    "model.add(Flatten())\n",
    "model.add(Dense(6, activation='softmax'))\n",
    "\n",
    "model.compile(loss='categorical_crossentropy', metrics=['accuracy'],optimizer='adam')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.fit(X, y_new, epochs=22, batch_size=64, shuffle=True, validation_split=0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.save('model.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import cv2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_img = cv2.imread('../input/happy-img-test/pexels-andrea-piacquadio-941693.jpg', 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_img.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_img = cv2.resize(test_img, (48,48))\n",
    "test_img.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_img = test_img.reshape(1,48,48,1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.predict(test_img)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# label_map = ['Anger', 'Neutral', 'Fear', 'Happy', 'Sad', 'Surprise']"
   ]
  }
 ],
 "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": "3.8.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}