as2-model-khera.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/importlib/_bootstrap.py:219: RuntimeWarning: compiletime version 3.6 of module 'tensorflow.python.framework.fast_tensor_util' does not match runtime version 3.7\n",
      "  return f(*args, **kwds)\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "import os,shutil\n",
    "import numpy as np\n",
    "import cv2\n",
    "import pickle\n",
    "import pandas as pd\n",
    "\n",
    "from tqdm import tqdm_notebook\n",
    "from matplotlib import pyplot as plt\n",
    "\n",
    "from sklearn.decomposition import PCA\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import classification_report\n",
    "from sklearn.metrics import f1_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Conv2D, Flatten, BatchNormalization, MaxPooling2D, Dropout, Activation\n",
    "from keras.initializers import glorot_normal\n",
    "from keras.optimizers import Adam\n",
    "from keras.callbacks import EarlyStopping\n",
    "from keras.losses import binary_crossentropy\n",
    "from keras.utils import to_categorical"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "cur_dir = os.getcwd()\n",
    "\n",
    "dataset_dir = os.path.join(cur_dir,\"Dataset\")\n",
    "\n",
    "labels_dir = os.path.join(dataset_dir,\"labels\")\n",
    "frames_dir = os.path.join(dataset_dir,\"frames\")\n",
    "\n",
    "n_lecs = len(os.listdir(labels_dir))\n",
    "\n",
    "listOfLectures = []\n",
    "\n",
    "for lec in sorted(os.listdir(frames_dir)):\n",
    "    if lec.startswith('.'):\n",
    "        continue\n",
    "    frame_list = []\n",
    "    lec_path = os.path.join(frames_dir,lec)\n",
    "    for frame in sorted(os.listdir(lec_path)):\n",
    "        frame_list.append([os.path.join(lec_path,frame),])\n",
    "    listOfLectures.append(frame_list)\n",
    "\n",
    "\n",
    "lec_num = 0\n",
    "for csv in sorted(os.listdir(labels_dir)):\n",
    "    csv_path = os.path.join(labels_dir,csv)\n",
    "    with open(csv_path,'r') as input:\n",
    "        data = input.read()\n",
    "        data = data.split()\n",
    "        for i in range(len(data)):\n",
    "            listOfLectures[lec_num][i].append(int(data[i]))\n",
    "    lec_num+=1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Window size is the size of window on either side. Using window_size = 2 will cause 5 frames to be stacked for one input example"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def load_data_fast(listOfLectures, resolution, window_size):\n",
    "    X_train = []\n",
    "    Y_train = []\n",
    "    for i in tqdm_notebook(range(len(listOfLectures))):\n",
    "        im_saved = []\n",
    "        for j in tqdm_notebook(range(window_size,len(listOfLectures[i])-window_size)):\n",
    "            if j == window_size:\n",
    "                im_saved = [cv2.resize(cv2.imread(listOfLectures[i][j+k][0], 0), resolution)\n",
    "                       for k in range(-window_size, window_size+1)]\n",
    "            else:\n",
    "                ims = im_saved[1:]\n",
    "                ims.append(cv2.resize(cv2.imread(listOfLectures[i][j+window_size][0], 0), resolution))\n",
    "                im_saved = ims\n",
    "            x = np.hstack(im_saved)\n",
    "            x = x[:,:,np.newaxis]\n",
    "            X_train.append(x)\n",
    "            Y_train.append(listOfLectures[i][j][1])\n",
    "    return X_train, Y_train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e020e4276b2746c6a31714dd53cef71e",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=35), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "bc5a4eecfd984af1bdef81dc1b19537d",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=525), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ae0e5c3397c54b5297b244ce22aa0170",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=754), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e6b00c0c0eed4a71a8c252dd2a36aec2",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=1030), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "fcab4604dfd547c0aee948c45e10ecda",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=981), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ae16cddfbed64992a9065949757098e9",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=850), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d678243da74d47d3adb277d0c3781495",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=865), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "7e8e1671a87a4eb584ddacee13232201",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=846), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "2f2108fa400d42c7906659d8129d4218",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=1135), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "6a80d713a04545d3b415aa43941e664b",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=783), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "8090d1959bd2475281cd9ca483d8aa8b",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=728), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "788c0679d1c34a2e81411e71932258ba",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=773), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "64da3f9de61f4092bf2aef2d193561ab",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=791), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "600bebd88c6e4a46a4613a863b721680",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=731), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "fc2cc09508da4a04a5ae0d182463c2d1",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=793), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e2ca17624a534699b3877ce486f787f1",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=409), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "b5cbb5b0ce2547f799eea58d06b2618d",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=475), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "4733740032a44e808ae488919a1dad09",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=777), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "7780712e69814cb99605d32293793b65",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=780), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "7a8965443a034baf964e33b76f8be87c",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=1128), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d4d2dccbbb3e4991b729eede1916bcff",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=862), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "29a783e8cf6c4a42a158bec86a0d0c60",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=924), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f1a6c93984464d77a55b2540e030e0ef",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=1116), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a222d8c1acc04d008a901adb9ba4d24d",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=400), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "100db863261c4520812e1a4de4e36e6d",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=841), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "fa193903c3854052add7638b056042e5",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=1040), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "9923b378e6e34d6c805ceaff3bc6289a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=1023), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a3634215d33445b2ad5df603206bf270",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=891), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "3a6728bf6898425696ff2b538c4885af",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=907), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "838da1157d044252ba350132d1cd02da",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=241), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "85aa3c57fc2844c5932c0ada3d006c4e",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=1107), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "27dbbf30072e4f4697fc291ee5a6ba85",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=862), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "cb33aa3a4771435a837364c7c4b8cc8a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=1002), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "c6faaa90a84c40f7a94307041c873d45",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=845), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "9993fbc478b5457e91493732bfc6b0f5",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=961), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "45ae6da9091c4eb4b5cd457ae27ed8d7",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=891), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "resolution = (60,80)\n",
    "window_size = 1\n",
    "X_train, Y_train = load_data_fast(listOfLectures, resolution, window_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(80, 180)\n",
      "(80, 180, 1)\n",
      "26160\n"
     ]
    }
   ],
   "source": [
    "print(X_train[0].shape)\n",
    "qwe = X_train[0][:, :, np.newaxis]\n",
    "print(qwe.shape)\n",
    "print(len(X_train))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(29067, 80, 180, 1)\n"
     ]
    }
   ],
   "source": [
    "X_train = np.stack(X_train)\n",
    "print(X_train.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Run the code below only if you change the architecture"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(\"./XY_train_80_60_3_horizontal_stack.pkl\", \"wb\") as pickle_in:\n",
    "    pickle.dump((X_train, Y_train), pickle_in)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(\"./XY_train_80_60_3_horizontal_stack.pkl\", \"rb\") as pickle_in:\n",
    "    X_train, Y_train = pickle.load(pickle_in)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "Y_train = to_categorical(Y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(26160, 2)\n",
      "(26160, 80, 180, 1)\n"
     ]
    }
   ],
   "source": [
    "print(Y_train.shape)\n",
    "print(X_train.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, Y_train, Y_test = train_test_split(X_train, Y_train, test_size=0.1, random_state=42, stratify=Y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(26160, 80, 180, 1)\n",
      "(2907, 80, 180, 1)\n"
     ]
    }
   ],
   "source": [
    "print(X_train.shape)\n",
    "print(X_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "n_zeros_train = np.sum(Y_train[:, 0])    # no. of 0s in Y_train\n",
    "n_ones_train = np.sum(Y_train[:, 1])    # no. of 1s in Y_train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'n_zeros_train' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-5-3111e976c57f>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mn_zeros_train\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m: name 'n_zeros_train' is not defined"
     ]
    }
   ],
   "source": [
    "n_zeros_train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.utils import resample\n",
    "scale_ones_to_zeros = 0.5\n",
    "X_train_pos = X_train[Y_train[:, 1] == 1]\n",
    "Y_train_pos = Y_train[Y_train[:, 1] == 1]\n",
    "X_train_ups, Y_train_ups =  resample(X_train_pos, Y_train_pos, \n",
    "                                 replace=True,     # sample with replacement\n",
    "                                 n_samples=int(n_zeros_train*scale_ones_to_zeros),    # to match majority class\n",
    "                                 random_state=123)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_final = np.concatenate((X_train[Y_train[:, 1] == 0], X_train_ups), axis=0)\n",
    "Y_train_final = np.concatenate((Y_train[Y_train[:, 1] == 0], Y_train_ups), axis=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(37762, 80, 180, 1)\n",
      "(37762, 2)\n"
     ]
    }
   ],
   "source": [
    "print(X_train_final.shape)\n",
    "print(Y_train_final.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "num_epochs = 5\n",
    "num_classes = 2\n",
    "# batch_size = 100\n",
    "learning_rate = 0.001"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [],
   "source": [
    "def gen_model():\n",
    "    model = Sequential()\n",
    "    model.add(Conv2D(64, kernel_size=3, activation='relu', input_shape= (80,180,1), padding='valid'))\n",
    "    model.add(Conv2D(64, kernel_size=3, activation='relu', padding='valid'))\n",
    "    model.add(BatchNormalization())\n",
    "    model.add(MaxPooling2D(pool_size=(2, 2), strides=None, padding='valid'))\n",
    "    \n",
    "    model.add(Conv2D(32, kernel_size=3, activation='relu'))\n",
    "    model.add(Conv2D(32, kernel_size=3, activation='relu'))\n",
    "    model.add(BatchNormalization())\n",
    "    model.add(MaxPooling2D(pool_size=(2, 2), strides=None, padding='valid'))\n",
    "    \n",
    "    model.add(Flatten())\n",
    "    model.add(Dense(2, activation='softmax'))\n",
    "    return model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = gen_model()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [],
   "source": [
    "optim = Adam(lr=learning_rate, beta_1=0.9, beta_2=0.999, amsgrad=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(loss='binary_crossentropy', optimizer=optim, metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [],
   "source": [
    "es = EarlyStopping(monitor='val_loss', mode='min', verbose=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 37762 samples, validate on 2907 samples\n",
      "Epoch 1/5\n",
      "37762/37762 [==============================] - 118s 3ms/step - loss: 0.6376 - acc: 0.8159 - val_loss: 1.2792 - val_acc: 0.7650\n",
      "Epoch 2/5\n",
      "37762/37762 [==============================] - 117s 3ms/step - loss: 0.3400 - acc: 0.8955 - val_loss: 0.2076 - val_acc: 0.9436\n",
      "Epoch 3/5\n",
      "26944/37762 [====================>.........] - ETA: 32s - loss: 0.2321 - acc: 0.9259"
     ]
    }
   ],
   "source": [
    "history = model.fit(X_train_final, Y_train_final, epochs=num_epochs, verbose=1, validation_data=(X_test, Y_test), callbacks=[es])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "_, train_acc = model.evaluate(X_train, Y_train, verbose=1)\n",
    "_, test_acc = model.evaluate(X_test, Y_test, verbose=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "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": [
    "plt.plot(history.history['loss'], label='train')\n",
    "plt.plot(history.history['val_loss'], label='test')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [],
   "source": [
    "preds = model.predict(X_test)\n",
    "preds = np.argmax(preds, axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.99      0.99      0.99      2798\n",
      "           1       0.84      0.86      0.85       109\n",
      "\n",
      "   micro avg       0.99      0.99      0.99      2907\n",
      "   macro avg       0.92      0.93      0.92      2907\n",
      "weighted avg       0.99      0.99      0.99      2907\n",
      "\n",
      "0.8506787330316743\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(np.argmax(Y_test, axis=1), preds))\n",
    "print(f1_score(np.argmax(Y_test, axis=1), preds))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.save('f1_0.8.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "im = cv2.imread(listOfLectures[0][1][0],0)\n",
    "hog = cv2.HOGDescriptor()\n",
    "h = hog.compute(im)\n",
    "plt.subplot(121),plt.imshow(im,cmap = 'gray')\n",
    "plt.title('Original Image'), plt.xticks([]), plt.yticks([])\n",
    "plt.subplot(122),plt.imshow(edges,cmap = 'gray')\n",
    "plt.title('Edge Image'), plt.xticks([]), plt.yticks([])\n",
    "plt.show()"
   ]
  },
  {
   "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": "3.7.0b3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}