fingerprintClassification_modelResNet_version2.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from tensorflow.keras.applications.resnet50 import preprocess_input, decode_predictions, ResNet50\n",
    "from tensorflow.keras.applications.resnet50 import ResNet50\n",
    "from tensorflow.keras import models\n",
    "from tensorflow.keras.models import Model\n",
    "from tensorflow.keras.layers import Input, Dense, Dropout\n",
    "from tensorflow.keras.layers import GlobalAveragePooling2D"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 3200 images belonging to 5 classes.\n",
      "Found 800 images belonging to 5 classes.\n"
     ]
    }
   ],
   "source": [
    "imagesize=256\n",
    "\n",
    "train_generator=ImageDataGenerator(featurewise_center=True, \n",
    "                                   vertical_flip=True,\n",
    "                                   preprocessing_function=preprocess_input)\n",
    "\n",
    "train_gen=train_generator.flow_from_directory('fingerprintClassificationDatabase/trainingSet',\n",
    "                                              target_size=(256,256),\n",
    "                                              batch_size=10,\n",
    "                                              class_mode='categorical')\n",
    "\n",
    "validation_generator=ImageDataGenerator(preprocessing_function=preprocess_input)\n",
    "\n",
    "validation_gen=validation_generator.flow_from_directory('fingerprintClassificationDatabase/testSet',\n",
    "                                                        target_size=(imagesize,imagesize),\n",
    "                                                        batch_size=10,\n",
    "                                                        class_mode='categorical')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "base_model = ResNet50(include_top=False, weights='imagenet')\n",
    "for layer in base_model.layers:\n",
    "    layer.trainable = True\n",
    "\n",
    "x = base_model.output\n",
    "x = GlobalAveragePooling2D()(x)\n",
    "\n",
    "x=Dense(512,activation='relu')(x)\n",
    "\n",
    "predictions = Dense(5, activation='softmax')(x)\n",
    "\n",
    "model = Model(inputs=base_model.input, outputs=predictions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.optimizers import Nadam\n",
    "model.compile(optimizer=Nadam(lr=0.0002, beta_1=0.9, beta_2=0.999, epsilon=None, schedule_decay=0.004), metrics=['accuracy'],loss='categorical_crossentropy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/15\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\mohib\\Anaconda3\\lib\\site-packages\\keras_preprocessing\\image.py:959: UserWarning: This ImageDataGenerator specifies `featurewise_center`, but it hasn't been fit on any training data. Fit it first by calling `.fit(numpy_data)`.\n",
      "  warnings.warn('This ImageDataGenerator specifies '\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "45/45 [==============================] - 801s 18s/step - loss: 1.1749 - acc: 0.5200 - val_loss: 6.6586 - val_acc: 0.3540\n",
      "Epoch 2/15\n",
      "45/45 [==============================] - 786s 17s/step - loss: 0.8135 - acc: 0.6956 - val_loss: 1.1603 - val_acc: 0.7220\n",
      "Epoch 3/15\n",
      "45/45 [==============================] - 785s 17s/step - loss: 0.6811 - acc: 0.7511 - val_loss: 0.5896 - val_acc: 0.7620\n",
      "Epoch 4/15\n",
      "45/45 [==============================] - 789s 18s/step - loss: 0.6469 - acc: 0.7889 - val_loss: 0.9061 - val_acc: 0.6760\n",
      "Epoch 5/15\n",
      "45/45 [==============================] - 787s 17s/step - loss: 0.5325 - acc: 0.8156 - val_loss: 0.4534 - val_acc: 0.8100\n",
      "Epoch 6/15\n",
      "45/45 [==============================] - 787s 17s/step - loss: 0.4439 - acc: 0.8311 - val_loss: 0.4050 - val_acc: 0.8440\n",
      "Epoch 7/15\n",
      "45/45 [==============================] - 787s 17s/step - loss: 0.4080 - acc: 0.8689 - val_loss: 0.4571 - val_acc: 0.8360\n",
      "Epoch 8/15\n",
      "45/45 [==============================] - 788s 18s/step - loss: 0.3941 - acc: 0.8689 - val_loss: 0.8888 - val_acc: 0.7100\n",
      "Epoch 9/15\n",
      "45/45 [==============================] - 798s 18s/step - loss: 0.3894 - acc: 0.8667 - val_loss: 0.5912 - val_acc: 0.7900\n",
      "Epoch 10/15\n",
      "45/45 [==============================] - 806s 18s/step - loss: 0.3961 - acc: 0.8667 - val_loss: 0.3272 - val_acc: 0.8800\n",
      "Epoch 11/15\n",
      "45/45 [==============================] - 816s 18s/step - loss: 0.3884 - acc: 0.8844 - val_loss: 0.4000 - val_acc: 0.8560\n",
      "Epoch 12/15\n",
      "45/45 [==============================] - 797s 18s/step - loss: 0.3273 - acc: 0.9022 - val_loss: 0.3463 - val_acc: 0.8840\n",
      "Epoch 13/15\n",
      "45/45 [==============================] - 789s 18s/step - loss: 0.3999 - acc: 0.8778 - val_loss: 0.3686 - val_acc: 0.8640\n",
      "Epoch 14/15\n",
      "45/45 [==============================] - 811s 18s/step - loss: 0.3776 - acc: 0.8800 - val_loss: 0.3594 - val_acc: 0.8760\n",
      "Epoch 15/15\n",
      "45/45 [==============================] - 811s 18s/step - loss: 0.2926 - acc: 0.9111 - val_loss: 0.4115 - val_acc: 0.8600\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x15fd099dac8>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit_generator(train_gen, \n",
    "                    steps_per_epoch=45,\n",
    "                    epochs=15,\n",
    "                    validation_data=validation_gen, \n",
    "                    validation_steps=50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss of trained model : 0.4010588064789772\n",
      "Accuracy of trained model : 0.8299999952316284\n"
     ]
    }
   ],
   "source": [
    "score_training=model.evaluate_generator(validation_gen, \n",
    "                         steps=10, \n",
    "                         verbose=0)\n",
    "print(\"Loss of trained model :\",score_training[0])\n",
    "print(\"Accuracy of trained model :\",score_training[1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(model.history.history['acc'])\n",
    "plt.plot(model.history.history['val_acc'])\n",
    "plt.title('Model Accuracy')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Training', 'Validation'], loc='center right')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(model.history.history['loss'])\n",
    "plt.plot(model.history.history['val_loss'])\n",
    "plt.title('Model loss')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Training', 'Validation'], loc='upper right')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.preprocessing import image \n",
    "test_image = image.load_img('left_loop.jpeg', target_size = (256, 256))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_image = image.img_to_array(test_image)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_image = np.expand_dims(test_image, axis = 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictions = model.predict(test_image)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Predictions:\n",
      "Arch: 0.22082084324210882 %\n",
      "Left Loop: 96.84082865715027 %\n",
      "Right Loop: 0.014203056343831122 %\n",
      "Tended Arch: 0.5413543432950974 %\n",
      "Whirl: 2.382793463766575 %\n"
     ]
    }
   ],
   "source": [
    "print(\"Predictions:\")\n",
    "print(\"Arch:\", (predictions[0][0])*100,\"%\")\n",
    "print(\"Left Loop:\", (predictions[0][1])*100,\"%\")\n",
    "print(\"Right Loop:\", (predictions[0][2])*100,\"%\")\n",
    "print(\"Tended Arch:\", (predictions[0][3])*100,\"%\")\n",
    "print(\"Whirl:\", (predictions[0][4])*100,\"%\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[2.2082084e-03, 9.6840829e-01, 1.4203056e-04, 5.4135434e-03,\n",
       "        2.3827935e-02]], dtype=float32)"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'A': 0, 'L': 1, 'R': 2, 'T': 3, 'W': 4}"
      ]
     },
     "execution_count": 87,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_gen.class_indices"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.models import load_model\n",
    "model.save('resNet_model_for_fingerprintClassification_pkl.pkl')  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.save('resNet_model_for_fingerprintClassification_hdf5.hd5')"
   ]
  },
  {
   "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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}