Covid_NonCovid_image_Classification.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "4007ba69",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import os\n",
    "from tqdm import tqdm\n",
    "import matplotlib.image as mpimg\n",
    "import random\n",
    "\n",
    "CATEGORIES = [\"COVID\", \"non-COVID\"]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "48c88c88",
   "metadata": {},
   "source": [
    "# Training dataset preparation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "b533a782",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 700/700 [00:00<00:00, 1905.82it/s]\n",
      "100%|██████████| 700/700 [00:00<00:00, 1977.67it/s]\n"
     ]
    }
   ],
   "source": [
    "training_data = []\n",
    "DATADIR = \"Train\"\n",
    "def create_training_data():\n",
    "    for category in CATEGORIES:  \n",
    "\n",
    "        path = os.path.join(DATADIR,category) \n",
    "        class_num = CATEGORIES.index(category)  \n",
    "\n",
    "        for img in tqdm(os.listdir(path)):  \n",
    "            try:\n",
    "                img_array = mpimg.imread(os.path.join(path,img)) \n",
    "               # img_array = img_array.resize(50, 50)\n",
    "                #print(img_array.shape)\n",
    "                training_data.append([img_array, class_num])\n",
    "                #print(img_array.shape)\n",
    "            except Exception as e:  \n",
    "                pass\n",
    "            \n",
    "\n",
    "create_training_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "81ee43a2",
   "metadata": {},
   "outputs": [],
   "source": [
    "random.shuffle(training_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "163547d6",
   "metadata": {},
   "outputs": [],
   "source": [
    "xtrain = []\n",
    "ytrain = []\n",
    "for features,label in training_data:\n",
    "    xtrain.append(features)\n",
    "    ytrain.append(label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "53cb315c",
   "metadata": {},
   "outputs": [],
   "source": [
    "xtrain = np.asarray(xtrain) \n",
    "ytrain = np.asarray(ytrain) \n",
    " "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "bf2e8724",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n"
     ]
    }
   ],
   "source": [
    "print(type(xtrain))\n",
    "print(type(ytrain))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "33e1c27c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "uint8\n"
     ]
    }
   ],
   "source": [
    "ytrain = ytrain.astype('uint8')\n",
    "print(ytrain.dtype)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "a0726c91",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1400, 100, 100)\n",
      "(1400,)\n"
     ]
    }
   ],
   "source": [
    "print(xtrain.shape)\n",
    "print(ytrain.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "17a47ef9",
   "metadata": {},
   "outputs": [],
   "source": [
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "38184bc1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1400, 100, 100, 1)\n"
     ]
    }
   ],
   "source": [
    "xtrain = xtrain.reshape(xtrain.shape[0],100,100,1)\n",
    "print(xtrain.shape)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "b5a20b02",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1. 0.]\n"
     ]
    }
   ],
   "source": [
    "from keras.utils import to_categorical\n",
    "ytrain =  to_categorical(ytrain)\n",
    "print(ytrain[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "427a60b9",
   "metadata": {},
   "source": [
    "# Test dataset preparatoin "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "172e906a",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 302/302 [00:00<00:00, 2257.35it/s]\n",
      "100%|██████████| 284/284 [00:00<00:00, 1863.88it/s]\n"
     ]
    }
   ],
   "source": [
    "testing_data = []\n",
    "DATADIR = \"Test\"\n",
    "\n",
    "def create_training_data():\n",
    "    for category in CATEGORIES:  \n",
    "\n",
    "        path = os.path.join(DATADIR,category) \n",
    "        class_num = CATEGORIES.index(category)  \n",
    "\n",
    "        for img in tqdm(os.listdir(path)):  \n",
    "            try:\n",
    "                img_array = mpimg.imread(os.path.join(path,img)) \n",
    "               # img_array = img_array.resize(50, 50)\n",
    "                #print(img_array.shape)\n",
    "                testing_data.append([img_array, class_num])\n",
    "                #print(img_array.shape)\n",
    "            except Exception as e:  \n",
    "                pass\n",
    "            \n",
    "\n",
    "create_training_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "60e71119",
   "metadata": {},
   "outputs": [],
   "source": [
    "random.shuffle(testing_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "a394ce6b",
   "metadata": {},
   "outputs": [],
   "source": [
    "xtest = []\n",
    "ytest = []\n",
    "for features,label in testing_data:\n",
    "    xtest.append(features)\n",
    "    ytest.append(label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "12a3d927",
   "metadata": {},
   "outputs": [],
   "source": [
    "xtest  = np.asarray(xtest) \n",
    "ytest  = np.asarray(ytest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "97cf442f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n"
     ]
    }
   ],
   "source": [
    "print(type(xtest))\n",
    "print(type(ytest))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "263a4e35",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "uint8\n"
     ]
    }
   ],
   "source": [
    "ytest = ytest.astype('uint8')\n",
    "print(ytest.dtype)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "15ebfde5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(586, 100, 100)\n",
      "(586,)\n"
     ]
    }
   ],
   "source": [
    "print(xtest.shape)\n",
    "print(ytest.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "8ea64f16",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(586, 100, 100, 1)\n"
     ]
    }
   ],
   "source": [
    "xtest =  xtest.reshape(xtest.shape[0],100,100,1)\n",
    "print(xtest.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "a8a66ded",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1. 0.]\n"
     ]
    }
   ],
   "source": [
    "ytest  =  to_categorical(ytest)\n",
    "print(ytest[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8d4fbe6b",
   "metadata": {},
   "source": [
    "# Normalization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "0770f6cb",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.0 0.0\n",
      "1.0 1.0\n",
      "0.0 0.0\n",
      "0.654997\n",
      "0.26767886\n",
      "1.2888691 1.2888691\n",
      "-2.446951 -2.446951\n"
     ]
    }
   ],
   "source": [
    "print (np.min(xtrain),np.min(xtest))\n",
    "\n",
    "print (np.max(ytrain),np.max(ytest)) \n",
    "print (np.min(ytrain),np.min(ytest))\n",
    "\n",
    "mean = np.mean(xtrain) \n",
    "std = np.std(xtrain)\n",
    "\n",
    "print(mean) \n",
    "print(std)\n",
    "\n",
    "xtrain -= mean \n",
    "xtrain /= std\n",
    "\n",
    "xtest -= mean \n",
    "xtest /= std \n",
    "print (np.max(xtrain),np.max(xtest))\n",
    "print (np.min(xtrain),np.min(xtest))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b0417a8e",
   "metadata": {},
   "source": [
    "# LeNet"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "01103439",
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.layers import Dropout\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "9c049a2a",
   "metadata": {},
   "outputs": [],
   "source": [
    "import keras\n",
    "from keras import models\n",
    "from keras import layers\n",
    "import tensorflow as tf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "feda0599",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d (Conv2D)              (None, 96, 96, 6)         156       \n",
      "_________________________________________________________________\n",
      "dropout (Dropout)            (None, 96, 96, 6)         0         \n",
      "_________________________________________________________________\n",
      "max_pooling2d (MaxPooling2D) (None, 48, 48, 6)         0         \n",
      "_________________________________________________________________\n",
      "conv2d_1 (Conv2D)            (None, 22, 22, 16)        2416      \n",
      "_________________________________________________________________\n",
      "max_pooling2d_1 (MaxPooling2 (None, 11, 11, 16)        0         \n",
      "_________________________________________________________________\n",
      "flatten (Flatten)            (None, 1936)              0         \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 120)               232440    \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 84)                10164     \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 2)                 170       \n",
      "=================================================================\n",
      "Total params: 245,346\n",
      "Trainable params: 245,346\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model = models.Sequential()\n",
    "\n",
    "model.add(layers.Conv2D(6, (5,5), input_shape=(100,100,1), activation='relu', strides=(1,1), padding='valid'))\n",
    "model.add(Dropout(0.2))\n",
    "\n",
    "model.add(layers.MaxPooling2D(pool_size=(2,2), strides=(2,2)))\n",
    "\n",
    "model.add(layers.Conv2D(16, (5,5),activation='relu', strides=(2,2), padding='valid'))\n",
    "\n",
    "\n",
    "model.add(layers.MaxPooling2D(pool_size=(2,2), strides=(2,2)))\n",
    "\n",
    "model.add(layers.Flatten())\n",
    "\n",
    "model.add(layers.Dense(120, activation='relu'))\n",
    "model.add(layers.Dense(84, activation='relu'))\n",
    "\n",
    "model.add(layers.Dense(2, activation='sigmoid'))\n",
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "98785509",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      "35/35 [==============================] - 6s 127ms/step - loss: 0.6552 - accuracy: 0.6180 - val_loss: 0.6198 - val_accuracy: 0.6500\n",
      "Epoch 2/50\n",
      "35/35 [==============================] - 4s 107ms/step - loss: 0.5360 - accuracy: 0.7479 - val_loss: 0.6715 - val_accuracy: 0.6143\n",
      "Epoch 3/50\n",
      "35/35 [==============================] - 4s 105ms/step - loss: 0.4843 - accuracy: 0.7696 - val_loss: 0.5866 - val_accuracy: 0.6893\n",
      "Epoch 4/50\n",
      "35/35 [==============================] - 4s 122ms/step - loss: 0.4199 - accuracy: 0.8241 - val_loss: 0.4587 - val_accuracy: 0.7929\n",
      "Epoch 5/50\n",
      "35/35 [==============================] - 6s 165ms/step - loss: 0.4262 - accuracy: 0.8073 - val_loss: 0.4914 - val_accuracy: 0.7500\n",
      "Epoch 6/50\n",
      "35/35 [==============================] - 5s 135ms/step - loss: 0.3928 - accuracy: 0.8319 - val_loss: 0.4001 - val_accuracy: 0.8250\n",
      "Epoch 7/50\n",
      "35/35 [==============================] - 4s 115ms/step - loss: 0.3458 - accuracy: 0.8525 - val_loss: 0.4226 - val_accuracy: 0.8000\n",
      "Epoch 8/50\n",
      "35/35 [==============================] - 4s 113ms/step - loss: 0.3740 - accuracy: 0.8233 - val_loss: 0.3866 - val_accuracy: 0.8321\n",
      "Epoch 9/50\n",
      "35/35 [==============================] - 4s 110ms/step - loss: 0.2938 - accuracy: 0.8818 - val_loss: 0.4163 - val_accuracy: 0.7857\n",
      "Epoch 10/50\n",
      "35/35 [==============================] - 4s 112ms/step - loss: 0.2841 - accuracy: 0.8836 - val_loss: 0.6188 - val_accuracy: 0.7000\n",
      "Epoch 11/50\n",
      "35/35 [==============================] - 4s 119ms/step - loss: 0.2883 - accuracy: 0.8752 - val_loss: 0.3529 - val_accuracy: 0.8393\n",
      "Epoch 12/50\n",
      "35/35 [==============================] - 4s 124ms/step - loss: 0.2534 - accuracy: 0.9002 - val_loss: 0.3341 - val_accuracy: 0.8500\n",
      "Epoch 13/50\n",
      "35/35 [==============================] - 4s 106ms/step - loss: 0.2105 - accuracy: 0.9249 - val_loss: 0.3299 - val_accuracy: 0.8643\n",
      "Epoch 14/50\n",
      "35/35 [==============================] - 4s 105ms/step - loss: 0.2282 - accuracy: 0.9124 - val_loss: 0.7021 - val_accuracy: 0.6571\n",
      "Epoch 15/50\n",
      "35/35 [==============================] - 4s 107ms/step - loss: 0.2940 - accuracy: 0.8728 - val_loss: 0.3331 - val_accuracy: 0.8393\n",
      "Epoch 16/50\n",
      "35/35 [==============================] - 4s 107ms/step - loss: 0.1947 - accuracy: 0.9252 - val_loss: 0.2983 - val_accuracy: 0.8571\n",
      "Epoch 17/50\n",
      "35/35 [==============================] - 4s 105ms/step - loss: 0.1744 - accuracy: 0.9326 - val_loss: 0.3611 - val_accuracy: 0.8214\n",
      "Epoch 18/50\n",
      "35/35 [==============================] - 4s 105ms/step - loss: 0.1645 - accuracy: 0.9371 - val_loss: 0.3504 - val_accuracy: 0.8286\n",
      "Epoch 19/50\n",
      "35/35 [==============================] - 4s 107ms/step - loss: 0.2342 - accuracy: 0.9089 - val_loss: 0.3296 - val_accuracy: 0.8393\n",
      "Epoch 20/50\n",
      "35/35 [==============================] - 4s 106ms/step - loss: 0.1683 - accuracy: 0.9327 - val_loss: 0.2772 - val_accuracy: 0.8607\n",
      "Epoch 21/50\n",
      "35/35 [==============================] - 4s 105ms/step - loss: 0.1421 - accuracy: 0.9556 - val_loss: 0.2802 - val_accuracy: 0.8893\n",
      "Epoch 22/50\n",
      "35/35 [==============================] - 4s 104ms/step - loss: 0.1510 - accuracy: 0.9401 - val_loss: 0.5917 - val_accuracy: 0.7643\n",
      "Epoch 23/50\n",
      "35/35 [==============================] - 4s 119ms/step - loss: 0.1519 - accuracy: 0.9413 - val_loss: 0.3615 - val_accuracy: 0.8357\n",
      "Epoch 24/50\n",
      "35/35 [==============================] - 4s 126ms/step - loss: 0.0873 - accuracy: 0.9715 - val_loss: 0.2899 - val_accuracy: 0.8571\n",
      "Epoch 25/50\n",
      "35/35 [==============================] - 4s 114ms/step - loss: 0.1155 - accuracy: 0.9671 - val_loss: 0.2846 - val_accuracy: 0.8679\n",
      "Epoch 26/50\n",
      "35/35 [==============================] - 4s 107ms/step - loss: 0.1018 - accuracy: 0.9696 - val_loss: 0.4683 - val_accuracy: 0.7714\n",
      "Epoch 27/50\n",
      "35/35 [==============================] - 4s 109ms/step - loss: 0.2185 - accuracy: 0.9247 - val_loss: 0.2741 - val_accuracy: 0.8964\n",
      "Epoch 28/50\n",
      "35/35 [==============================] - 4s 120ms/step - loss: 0.0827 - accuracy: 0.9782 - val_loss: 0.2550 - val_accuracy: 0.9036\n",
      "Epoch 29/50\n",
      "35/35 [==============================] - 4s 110ms/step - loss: 0.0642 - accuracy: 0.9797 - val_loss: 0.2520 - val_accuracy: 0.9036\n",
      "Epoch 30/50\n",
      "35/35 [==============================] - 5s 138ms/step - loss: 0.0583 - accuracy: 0.9844 - val_loss: 0.2694 - val_accuracy: 0.8821\n",
      "Epoch 31/50\n",
      "35/35 [==============================] - 5s 132ms/step - loss: 0.0751 - accuracy: 0.9806 - val_loss: 0.2648 - val_accuracy: 0.8821\n",
      "Epoch 32/50\n",
      "35/35 [==============================] - 4s 106ms/step - loss: 0.0765 - accuracy: 0.9945 - val_loss: 0.2619 - val_accuracy: 0.9107\n",
      "Epoch 33/50\n",
      "35/35 [==============================] - 5s 122ms/step - loss: 0.0441 - accuracy: 0.9940 - val_loss: 0.2599 - val_accuracy: 0.9036\n",
      "Epoch 34/50\n",
      "35/35 [==============================] - 5s 131ms/step - loss: 0.0424 - accuracy: 0.9940 - val_loss: 0.2618 - val_accuracy: 0.9036\n",
      "Epoch 35/50\n",
      "35/35 [==============================] - 5s 131ms/step - loss: 0.0305 - accuracy: 0.9978 - val_loss: 0.2570 - val_accuracy: 0.9143\n",
      "Epoch 36/50\n",
      "35/35 [==============================] - 4s 115ms/step - loss: 0.0268 - accuracy: 0.9951 - val_loss: 0.2502 - val_accuracy: 0.9143\n",
      "Epoch 37/50\n",
      "35/35 [==============================] - 4s 128ms/step - loss: 0.0211 - accuracy: 0.9975 - val_loss: 0.2659 - val_accuracy: 0.9036\n",
      "Epoch 38/50\n",
      "35/35 [==============================] - 5s 136ms/step - loss: 0.0218 - accuracy: 0.9985 - val_loss: 0.2667 - val_accuracy: 0.9107\n",
      "Epoch 39/50\n",
      "35/35 [==============================] - 5s 146ms/step - loss: 0.0186 - accuracy: 0.9980 - val_loss: 0.2599 - val_accuracy: 0.9036\n",
      "Epoch 40/50\n",
      "35/35 [==============================] - 4s 127ms/step - loss: 0.0151 - accuracy: 1.0000 - val_loss: 0.2503 - val_accuracy: 0.9214\n",
      "Epoch 41/50\n",
      "35/35 [==============================] - 5s 139ms/step - loss: 0.0139 - accuracy: 1.0000 - val_loss: 0.2579 - val_accuracy: 0.9179\n",
      "Epoch 42/50\n",
      "35/35 [==============================] - 5s 154ms/step - loss: 0.0099 - accuracy: 1.0000 - val_loss: 0.2645 - val_accuracy: 0.9179\n",
      "Epoch 43/50\n",
      "35/35 [==============================] - 6s 171ms/step - loss: 0.0105 - accuracy: 0.9997 - val_loss: 0.2608 - val_accuracy: 0.9143\n",
      "Epoch 44/50\n",
      "35/35 [==============================] - 6s 177ms/step - loss: 0.0090 - accuracy: 0.9999 - val_loss: 0.2812 - val_accuracy: 0.9036\n",
      "Epoch 45/50\n",
      "35/35 [==============================] - 5s 138ms/step - loss: 0.0084 - accuracy: 1.0000 - val_loss: 0.2632 - val_accuracy: 0.9179\n",
      "Epoch 46/50\n",
      "35/35 [==============================] - 5s 142ms/step - loss: 0.0079 - accuracy: 1.0000 - val_loss: 0.2739 - val_accuracy: 0.9107\n",
      "Epoch 47/50\n",
      "35/35 [==============================] - 6s 159ms/step - loss: 0.0063 - accuracy: 1.0000 - val_loss: 0.2678 - val_accuracy: 0.9179\n",
      "Epoch 48/50\n",
      "35/35 [==============================] - 5s 136ms/step - loss: 0.0060 - accuracy: 1.0000 - val_loss: 0.2712 - val_accuracy: 0.9214\n",
      "Epoch 49/50\n",
      "35/35 [==============================] - 5s 132ms/step - loss: 0.0070 - accuracy: 1.0000 - val_loss: 0.2696 - val_accuracy: 0.9214\n",
      "Epoch 50/50\n",
      "35/35 [==============================] - 5s 138ms/step - loss: 0.0061 - accuracy: 1.0000 - val_loss: 0.2682 - val_accuracy: 0.9179\n"
     ]
    }
   ],
   "source": [
    "model.compile(loss=\"categorical_crossentropy\", optimizer=\"sgd\", metrics=['accuracy'])\n",
    "history=model.fit(xtrain, ytrain, validation_split=(0.2), batch_size=32,epochs=50, verbose=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a839e6ab",
   "metadata": {},
   "source": [
    "# Evaluation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "6d9833e1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "19/19 [==============================] - 1s 33ms/step - loss: 1.7792 - accuracy: 0.6041\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[1.7791595458984375, 0.6040955781936646]"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.evaluate(xtest,ytest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "ce45c94c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "19/19 [==============================] - 1s 33ms/step - loss: 1.7792 - accuracy: 0.6041\n",
      "1.7791595458984375\n",
      "0.6040955781936646\n"
     ]
    }
   ],
   "source": [
    "test_loss,test_acc=model.evaluate(xtest,ytest)\n",
    "print(test_loss)\n",
    "print(test_acc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "59d09474",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_loss = history.history['loss']\n",
    "train_acc = history.history['accuracy']\n",
    "val_loss = history.history['val_loss']\n",
    "val_acc = history.history['val_accuracy']\n",
    "xc=range(len(train_loss))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "49833e96",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7fb15403b460>"
      ]
     },
     "execution_count": 54,
     "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"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "plt.plot(xc, train_loss)\n",
    "plt.plot(xc, val_loss)\n",
    "plt.xlabel(\"Epoch\")\n",
    "plt.ylabel(\"Loss\")\n",
    "plt.legend(['train','val'],loc=4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "b7bae976",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7fb13443ea90>"
      ]
     },
     "execution_count": 55,
     "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"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "plt.plot(xc, train_acc)\n",
    "plt.plot(xc, val_acc)\n",
    "plt.xlabel(\"Epoch\")\n",
    "plt.ylabel(\"Accuracy\")\n",
    "plt.legend(['train','val'],loc=4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "36fcdae8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1 Actual is  [1. 0.] Predicted is 1\n",
      "2 Actual is  [0. 1.] Predicted is 0\n",
      "3 Actual is  [0. 1.] Predicted is 1\n",
      "4 Actual is  [1. 0.] Predicted is 1\n",
      "5 Actual is  [1. 0.] Predicted is 0\n",
      "6 Actual is  [0. 1.] Predicted is 1\n",
      "7 Actual is  [1. 0.] Predicted is 1\n",
      "8 Actual is  [0. 1.] Predicted is 1\n",
      "9 Actual is  [1. 0.] Predicted is 1\n",
      "10 Actual is  [0. 1.] Predicted is 1\n"
     ]
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "for i in range(10):\n",
    "    pred = model.predict(xtest[i].reshape(1, 100, 100, 1))\n",
    "    print(\"\"+str(i+1)+\" Actual is \"+ \" \" + str(ytest[i])+\" Predicted is \"+str(pred.argmax()))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ce3013a1",
   "metadata": {},
   "source": [
    "# Accuracy, precision, recall ,f1 score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "id": "a4057a21",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.604096\n",
      "Precision: 0.561321\n",
      "Recall: 0.838028\n",
      "F1 score: 0.672316\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import accuracy_score\n",
    "from sklearn.metrics import precision_score\n",
    "from sklearn.metrics import recall_score\n",
    "from sklearn.metrics import f1_score\n",
    "\n",
    "accuracy = accuracy_score(ytest1,y_pred)\n",
    "print('Accuracy: %f' % accuracy)\n",
    "\n",
    "precision = precision_score(ytest1,y_pred)\n",
    "print('Precision: %f' % precision)\n",
    "\n",
    "recall = recall_score(ytest1,y_pred)\n",
    "print('Recall: %f' % recall)\n",
    "\n",
    "f1 = f1_score(ytest1,y_pred)\n",
    "print('F1 score: %f' % f1)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a57ec736",
   "metadata": {},
   "source": [
    "# confusion matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "id": "a402fc55",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "confusion matrix : \n",
      " [[116 186]\n",
      " [ 46 238]]\n"
     ]
    }
   ],
   "source": [
    "\n",
    "from sklearn.metrics import confusion_matrix,classification_report\n",
    "ytest1 = np.argmax(np.array(ytest), axis=1)\n",
    "predictions = model.predict(xtest)\n",
    "y_pred = np.argmax(predictions, axis=1)\n",
    "\n",
    "cm=confusion_matrix(ytest1,y_pred)\n",
    "print(\"confusion matrix : \\n\" ,cm)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "009b50f4",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f5b8229a",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b4cc3f4c",
   "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": 5
}