Konvolutivna_mreza_jednostavna.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Ucitavanje 5744 slika je trajalo 6 s\n",
      "Ucitavanje 1919 slika je trajalo 2 s\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "from keras import *\n",
    "import tensorflow as tf\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Dropout, Activation, Flatten\n",
    "from keras.optimizers import Adam\n",
    "from keras.layers.normalization import BatchNormalization\n",
    "from keras.utils import np_utils\n",
    "from keras.layers import Conv2D, MaxPooling2D, ZeroPadding2D, GlobalAveragePooling2D\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn import datasets, svm, metrics\n",
    "from skimage import io\n",
    "import os\n",
    "import matplotlib.pyplot\n",
    "from matplotlib.pyplot import imread\n",
    "import skimage.transform\n",
    "from skimage.transform import resize as imresize\n",
    "from sklearn.model_selection import train_test_split\n",
    "from keras import regularizers\n",
    "import csv\n",
    "import time\n",
    "import cv2\n",
    "\n",
    "IMAGE_WIDTH = 32\n",
    "IMAGE_HEIGHT = 32\n",
    "IMAGE_DIMENSIONS = (IMAGE_HEIGHT, IMAGE_WIDTH)\n",
    "\n",
    "TRENING_PUTANJA = './fruits-360/Train'\n",
    "VALIDACIONA_PUTANJA = './fruits-360/Test'\n",
    "\n",
    "# kao parametar prima putanju do folder-a dataset-a\n",
    "def ucitajSlikeIzFoldera(putanja):\n",
    "    slike = []\n",
    "    labele = []\n",
    "    idx_mapa = []\n",
    "    idx = -1\n",
    "\n",
    "    subfolderi = os.listdir(putanja)\n",
    "\n",
    "    for naziv_klase in subfolderi:\n",
    "        putanja_do_slika = os.path.join(putanja, naziv_klase)\n",
    "        labele.append(naziv_klase)\n",
    "\n",
    "        idx = idx + 1\n",
    "        \n",
    "        for naziv_slike in os.listdir(putanja_do_slika):\n",
    "            putanja_do_slike = os.path.join(putanja_do_slika, naziv_slike)\n",
    "            slika = cv2.imread(putanja_do_slike)\n",
    "            slika = cv2.resize(slika, IMAGE_DIMENSIONS)\n",
    "            slika = cv2.cvtColor(slika, cv2.COLOR_RGB2BGR)\n",
    "            slike.append(slika)\n",
    "            \n",
    "            #punim array koji mapira sliku na index labele\n",
    "            idx_mapa.append(idx)\n",
    "            \n",
    "    return np.array(slike), np.array(idx_mapa), labele\n",
    "\n",
    "\n",
    "start = time.time()\n",
    "X_train, y_train, labele = ucitajSlikeIzFoldera(TRENING_PUTANJA)\n",
    "end = time.time()\n",
    "print(\"Ucitavanje %d slika je trajalo %d s\" % (len(y_train),end-start))\n",
    "\n",
    "start = time.time()\n",
    "X_test, y_test, unused = ucitajSlikeIzFoldera(VALIDACIONA_PUTANJA)\n",
    "end = time.time()\n",
    "print(\"Ucitavanje %d slika je trajalo %d s\" % (len(y_test),end-start))\n",
    "\n",
    "X_val, X_test , y_val, y_test = train_test_split(X_test, y_test, test_size = 0.66, random_state = 0)\n",
    "\n",
    "# ukupno = len(y_train) + len(y_test) + len(y_val)\n",
    "# print(\"Velicina trening set-a: %d \" % (len(y_train) * 100 / ukupno))\n",
    "# print(\"Velicina validacionog set-a: %d \" % (len(y_val) * 100 / ukupno))\n",
    "# print(\"Velicina test set-a: %d \" % (len(y_test) * 100 / ukupno))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "Y_train = np_utils.to_categorical(y_train, 15)\n",
    "Y_test = np_utils.to_categorical(y_test, 15)\n",
    "Y_val = np_utils.to_categorical(y_val, 15)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 5744 samples, validate on 652 samples\n",
      "Epoch 1/10\n",
      "5744/5744 [==============================] - 43s 7ms/step - loss: 0.6864 - acc: 0.8341 - val_loss: 0.0819 - val_acc: 0.9678\n",
      "Epoch 2/10\n",
      "5744/5744 [==============================] - 42s 7ms/step - loss: 0.0243 - acc: 0.9929 - val_loss: 0.0722 - val_acc: 0.9755\n",
      "Epoch 3/10\n",
      "5744/5744 [==============================] - 45s 8ms/step - loss: 0.0391 - acc: 0.9864 - val_loss: 0.0938 - val_acc: 0.9678\n",
      "Epoch 4/10\n",
      "5744/5744 [==============================] - 46s 8ms/step - loss: 0.0143 - acc: 0.9958 - val_loss: 0.2014 - val_acc: 0.9172\n",
      "Epoch 5/10\n",
      "5744/5744 [==============================] - 42s 7ms/step - loss: 0.0050 - acc: 0.9995 - val_loss: 0.0579 - val_acc: 0.9877\n",
      "Epoch 6/10\n",
      "5744/5744 [==============================] - 41s 7ms/step - loss: 0.0038 - acc: 0.9993 - val_loss: 0.0209 - val_acc: 0.9893\n",
      "Epoch 7/10\n",
      "5744/5744 [==============================] - 39s 7ms/step - loss: 0.0061 - acc: 0.9979 - val_loss: 0.1296 - val_acc: 0.9402\n",
      "Epoch 8/10\n",
      "5744/5744 [==============================] - 38s 7ms/step - loss: 0.0029 - acc: 0.9995 - val_loss: 0.0296 - val_acc: 0.9877\n",
      "Epoch 9/10\n",
      "5744/5744 [==============================] - 40s 7ms/step - loss: 0.0023 - acc: 0.9991 - val_loss: 0.0091 - val_acc: 0.9985\n",
      "Epoch 10/10\n",
      "5744/5744 [==============================] - 39s 7ms/step - loss: 7.4231e-04 - acc: 1.0000 - val_loss: 0.0400 - val_acc: 0.9801\n"
     ]
    }
   ],
   "source": [
    "tf.disable_v2_behavior()\n",
    "\n",
    "model = Sequential()\n",
    "\n",
    "model.add(Conv2D(32,(3,3),input_shape=(32,32,3), padding='same'))\n",
    "model.add(Activation('relu'))\n",
    "model.add(BatchNormalization())\n",
    "model.add(MaxPooling2D(pool_size=(2,2)))\n",
    "\n",
    "model.add(Conv2D(64,(3,3), padding='same'))\n",
    "model.add(Activation('relu'))\n",
    "model.add(BatchNormalization())\n",
    "model.add(MaxPooling2D(pool_size=(2,2)))\n",
    "model.add(Dropout(0.25))\n",
    "\n",
    "\n",
    "model.add(Conv2D(128,(3,3), padding='same'))\n",
    "model.add(Activation('relu'))\n",
    "model.add(BatchNormalization())\n",
    "model.add(MaxPooling2D(pool_size=(2,2)))\n",
    "\n",
    "model.add(Flatten())\n",
    "model.add(Dense(256))\n",
    "model.add(Activation('relu'))\n",
    "model.add(Dropout(0.2))\n",
    "model.add(Dense(12))\n",
    "model.add(Activation('softmax'))\n",
    "\n",
    "model.compile(loss='categorical_crossentropy', optimizer=Adam(),metrics=['accuracy'])\n",
    "history = model.fit(X_train,Y_train, epochs=10, batch_size = 128, validation_data=(X_val,Y_val))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1267/1267 [==============================] - 3s 2ms/step\n",
      "\n",
      "Test loss:  0.030450728942802575\n",
      "Test Accuracy 0.9881610105897652\n"
     ]
    }
   ],
   "source": [
    "score = model.evaluate(X_test, Y_test)\n",
    "print()\n",
    "print('Test loss: ', score[0])\n",
    "print('Test Accuracy', score[1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "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"
    },
    {
     "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['acc'])\n",
    "plt.plot(history.history['val_acc'])\n",
    "plt.title('model accuracy')\n",
    "plt.ylabel('accuracy')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['train', 'test'], loc='upper left')\n",
    "plt.show()\n",
    "\n",
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.title('model loss')\n",
    "plt.ylabel('loss')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['train', 'test'], loc='upper left')\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": "3.5.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}