ASL_train.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\ryans\\Anaconda3\\lib\\site-packages\\numpy\\_distributor_init.py:32: UserWarning: loaded more than 1 DLL from .libs:\n",
      "C:\\Users\\ryans\\Anaconda3\\lib\\site-packages\\numpy\\.libs\\libopenblas.NOIJJG62EMASZI6NYURL6JBKM4EVBGM7.gfortran-win_amd64.dll\n",
      "C:\\Users\\ryans\\Anaconda3\\lib\\site-packages\\numpy\\.libs\\libopenblas.PYQHXLVVQ7VESDPUVUADXEVJOBGHJPAY.gfortran-win_amd64.dll\n",
      "  stacklevel=1)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'test': 0, 'train': 1}\n",
      "['test', 'train']\n",
      "[0, 1]\n"
     ]
    }
   ],
   "source": [
    "import cv2,os\n",
    "\n",
    "data_path='DATASET'\n",
    "categories=os.listdir(data_path)\n",
    "labels=[i for i in range(len(categories))]\n",
    "\n",
    "label_dict=dict(zip(categories,labels)) #empty dictionary\n",
    "\n",
    "print(label_dict)\n",
    "print(categories)\n",
    "print(labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['0', 'A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 'J', 'K', 'L', 'M', 'N', 'O', 'P', 'Q', 'R', 'S', 'T', 'U', 'V', 'W', 'X', 'Y', 'Z']\n",
      "[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26]\n"
     ]
    }
   ],
   "source": [
    "data_path='DATASET/train'\n",
    "classes_path=os.listdir(data_path)\n",
    "classesf=os.listdir(data_path)\n",
    "print(classesf)\n",
    "labels_classes=[i for i in range(len(classesf))]\n",
    "print(labels_classes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_path='DATASET'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [],
   "source": [
    "label_classes_dict=dict(zip(classesf,labels_classes))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'0': 0, 'A': 1, 'B': 2, 'C': 3, 'D': 4, 'E': 5, 'F': 6, 'G': 7, 'H': 8, 'I': 9, 'J': 10, 'K': 11, 'L': 12, 'M': 13, 'N': 14, 'O': 15, 'P': 16, 'Q': 17, 'R': 18, 'S': 19, 'T': 20, 'U': 21, 'V': 22, 'W': 23, 'X': 24, 'Y': 25, 'Z': 26}\n"
     ]
    }
   ],
   "source": [
    "#print(labels_classes)\n",
    "#print(categories)\n",
    "print(label_classes_dict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DATASET\\test\n",
      "['0', 'A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 'J', 'K', 'L', 'M', 'N', 'O', 'P', 'Q', 'R', 'S', 'T', 'U', 'V', 'W', 'X', 'Y', 'Z']\n",
      "DATASET\\test\\0\n",
      "DATASET\\test\\A\n",
      "DATASET\\test\\B\n",
      "DATASET\\test\\C\n",
      "DATASET\\test\\D\n",
      "DATASET\\test\\E\n",
      "DATASET\\test\\F\n",
      "DATASET\\test\\G\n",
      "DATASET\\test\\H\n",
      "DATASET\\test\\I\n",
      "DATASET\\test\\J\n",
      "DATASET\\test\\K\n",
      "DATASET\\test\\L\n",
      "DATASET\\test\\M\n",
      "DATASET\\test\\N\n",
      "DATASET\\test\\O\n",
      "DATASET\\test\\P\n",
      "DATASET\\test\\Q\n",
      "DATASET\\test\\R\n",
      "DATASET\\test\\S\n",
      "DATASET\\test\\T\n",
      "DATASET\\test\\U\n",
      "DATASET\\test\\V\n",
      "DATASET\\test\\W\n",
      "DATASET\\test\\X\n",
      "DATASET\\test\\Y\n",
      "DATASET\\test\\Z\n",
      "DATASET\\train\n",
      "['0', 'A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 'J', 'K', 'L', 'M', 'N', 'O', 'P', 'Q', 'R', 'S', 'T', 'U', 'V', 'W', 'X', 'Y', 'Z']\n",
      "DATASET\\train\\0\n",
      "DATASET\\train\\A\n",
      "DATASET\\train\\B\n",
      "DATASET\\train\\C\n",
      "DATASET\\train\\D\n",
      "DATASET\\train\\E\n",
      "DATASET\\train\\F\n",
      "DATASET\\train\\G\n",
      "DATASET\\train\\H\n",
      "DATASET\\train\\I\n",
      "DATASET\\train\\J\n",
      "DATASET\\train\\K\n",
      "DATASET\\train\\L\n",
      "DATASET\\train\\M\n",
      "DATASET\\train\\N\n",
      "DATASET\\train\\O\n",
      "DATASET\\train\\P\n",
      "DATASET\\train\\Q\n",
      "DATASET\\train\\R\n",
      "DATASET\\train\\S\n",
      "DATASET\\train\\T\n",
      "DATASET\\train\\U\n",
      "DATASET\\train\\V\n",
      "DATASET\\train\\W\n",
      "DATASET\\train\\X\n",
      "DATASET\\train\\Y\n",
      "DATASET\\train\\Z\n"
     ]
    }
   ],
   "source": [
    "img_size=128\n",
    "data=[]\n",
    "target=[]\n",
    "c=0\n",
    "minValue = 70\n",
    "for category in categories:\n",
    "    \n",
    "    cat_path=os.path.join(data_path,category)\n",
    "    print(cat_path)\n",
    "    cat_names=os.listdir(cat_path)\n",
    "    print(cat_names)\n",
    "    for classes in cat_names:\n",
    "        folder_path=os.path.join(data_path,category,classes)\n",
    "        print(folder_path)\n",
    "        img_names=os.listdir(folder_path)\n",
    "        #print(img_names)\n",
    "        for img_name in img_names:\n",
    "            #print(img_name)\n",
    "            img_path=os.path.join(folder_path,img_name)\n",
    "            img=cv2.imread(img_path)\n",
    "            \n",
    "            try:\n",
    "                gray=cv2.cvtColor(img,cv2.COLOR_BGR2GRAY) \n",
    "                blur = cv2.GaussianBlur(gray,(5,5),2)\n",
    "                th3 = cv2.adaptiveThreshold(blur,255,cv2.ADAPTIVE_THRESH_GAUSSIAN_C,cv2.THRESH_BINARY_INV,11,2)\n",
    "                ret, res = cv2.threshold(th3, minValue, 255, cv2.THRESH_BINARY_INV+cv2.THRESH_OTSU)\n",
    "                #res=np.array(res)\n",
    "                #print(type(res))\n",
    "            #Converting the image into gray scale\n",
    "                resized=cv2.resize(res,(img_size,img_size))\n",
    "            #resizing the gray scale into 50x50, since we need a fixed common size for all the images in the dataset\n",
    "                data.append(resized)\n",
    "                #print(data)\n",
    "                target.append(label_classes_dict[classes])\n",
    "            except Exception as e:\n",
    "                print('Exception:',e)\n",
    "            \n",
    "            \n",
    "        \n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [],
   "source": [
    "datanp=np.array(data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(17113, 128, 128)"
      ]
     },
     "execution_count": 84,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "datanp.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(17113,)"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "targetnp=np.array(target)\n",
    "\n",
    "targetnp.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "\n",
    "data=np.array(data)/255.0\n",
    "data=np.reshape(data,(data.shape[0],img_size,img_size,1))\n",
    "target=np.array(target)\n",
    "\n",
    "from keras.utils import np_utils\n",
    "\n",
    "new_target=np_utils.to_categorical(target)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(17113, 27)"
      ]
     },
     "execution_count": 87,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "new_target.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.save('data_img',data)\n",
    "np.save('target',new_target)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [],
   "source": [
    "data=np.load('data_img.npy')\n",
    "target=np.load('target.npy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "train_data,test_data,train_target,test_target=train_test_split(data,new_target,test_size=0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_2\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d_3 (Conv2D)            (None, 126, 126, 32)      320       \n",
      "_________________________________________________________________\n",
      "max_pooling2d_3 (MaxPooling2 (None, 63, 63, 32)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_4 (Conv2D)            (None, 61, 61, 32)        9248      \n",
      "_________________________________________________________________\n",
      "max_pooling2d_4 (MaxPooling2 (None, 30, 30, 32)        0         \n",
      "_________________________________________________________________\n",
      "flatten_2 (Flatten)          (None, 28800)             0         \n",
      "_________________________________________________________________\n",
      "dense_5 (Dense)              (None, 128)               3686528   \n",
      "_________________________________________________________________\n",
      "dropout_3 (Dropout)          (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "dense_6 (Dense)              (None, 96)                12384     \n",
      "_________________________________________________________________\n",
      "dropout_4 (Dropout)          (None, 96)                0         \n",
      "_________________________________________________________________\n",
      "dense_7 (Dense)              (None, 64)                6208      \n",
      "_________________________________________________________________\n",
      "dense_8 (Dense)              (None, 27)                1755      \n",
      "=================================================================\n",
      "Total params: 3,716,443\n",
      "Trainable params: 3,716,443\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Convolution2D\n",
    "from keras.layers import MaxPooling2D\n",
    "from keras.layers import Flatten\n",
    "from keras.layers import Dense , Dropout\n",
    "import os\n",
    "os.environ[\"CUDA_VISIBLE_DEVICES\"] = \"1\"\n",
    "sz = 128\n",
    "# Step 1 - Building the CNN\n",
    "\n",
    "# Initializing the CNN\n",
    "classifier = Sequential()\n",
    "\n",
    "# First convolution layer and pooling\n",
    "classifier.add(Convolution2D(32, (3, 3), input_shape=(sz, sz, 1), activation='relu'))\n",
    "classifier.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "# Second convolution layer and pooling\n",
    "classifier.add(Convolution2D(32, (3, 3), activation='relu'))\n",
    "# input_shape is going to be the pooled feature maps from the previous convolution layer\n",
    "classifier.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "#classifier.add(Convolution2D(32, (3, 3), activation='relu'))\n",
    "# input_shape is going to be the pooled feature maps from the previous convolution layer\n",
    "#classifier.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "\n",
    "# Flattening the layers\n",
    "classifier.add(Flatten())\n",
    "\n",
    "# Adding a fully connected layer\n",
    "classifier.add(Dense(units=128, activation='relu'))\n",
    "classifier.add(Dropout(0.40))\n",
    "classifier.add(Dense(units=96, activation='relu'))\n",
    "classifier.add(Dropout(0.40))\n",
    "classifier.add(Dense(units=64, activation='relu'))\n",
    "classifier.add(Dense(units=27, activation='softmax')) # softmax for more than 2\n",
    "\n",
    "# Compiling the CNN\n",
    "classifier.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy']) # categorical_crossentropy for more than 2\n",
    "\n",
    "\n",
    "# Step 2 - Preparing the train/test data and training the model\n",
    "classifier.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.callbacks import ModelCheckpoint"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 9583 samples, validate on 4107 samples\n",
      "Epoch 1/20\n",
      "9583/9583 [==============================] - 108s 11ms/step - loss: 2.3023 - accuracy: 0.3101 - val_loss: 0.5760 - val_accuracy: 0.8795\n",
      "Epoch 2/20\n",
      "9583/9583 [==============================] - 98s 10ms/step - loss: 0.7736 - accuracy: 0.7359 - val_loss: 0.1293 - val_accuracy: 0.9827\n",
      "Epoch 3/20\n",
      "9583/9583 [==============================] - 96s 10ms/step - loss: 0.4438 - accuracy: 0.8519 - val_loss: 0.0425 - val_accuracy: 0.9905\n",
      "Epoch 4/20\n",
      "9583/9583 [==============================] - 90s 9ms/step - loss: 0.3134 - accuracy: 0.8954 - val_loss: 0.0172 - val_accuracy: 0.9971\n",
      "Epoch 5/20\n",
      "9583/9583 [==============================] - 91s 10ms/step - loss: 0.2613 - accuracy: 0.9138 - val_loss: 0.0131 - val_accuracy: 0.9973\n",
      "Epoch 6/20\n",
      "9583/9583 [==============================] - 91s 9ms/step - loss: 0.2097 - accuracy: 0.9323 - val_loss: 0.0155 - val_accuracy: 0.9966\n",
      "Epoch 7/20\n",
      "9583/9583 [==============================] - 90s 9ms/step - loss: 0.1818 - accuracy: 0.9375 - val_loss: 0.0070 - val_accuracy: 0.9985\n",
      "Epoch 8/20\n",
      "9583/9583 [==============================] - 91s 9ms/step - loss: 0.1720 - accuracy: 0.9469 - val_loss: 0.0062 - val_accuracy: 0.9985\n",
      "Epoch 9/20\n",
      "9583/9583 [==============================] - 91s 9ms/step - loss: 0.1535 - accuracy: 0.9506 - val_loss: 0.0093 - val_accuracy: 0.9978\n",
      "Epoch 10/20\n",
      "9583/9583 [==============================] - 91s 10ms/step - loss: 0.1428 - accuracy: 0.9562 - val_loss: 0.0064 - val_accuracy: 0.9983\n",
      "Epoch 11/20\n",
      "9583/9583 [==============================] - 94s 10ms/step - loss: 0.1251 - accuracy: 0.9590 - val_loss: 0.0059 - val_accuracy: 0.9985\n",
      "Epoch 12/20\n",
      "9583/9583 [==============================] - 90s 9ms/step - loss: 0.1283 - accuracy: 0.9624 - val_loss: 0.0048 - val_accuracy: 0.9993\n",
      "Epoch 13/20\n",
      "9583/9583 [==============================] - 91s 9ms/step - loss: 0.1034 - accuracy: 0.9691 - val_loss: 0.0087 - val_accuracy: 0.9976\n",
      "Epoch 14/20\n",
      "9583/9583 [==============================] - 91s 10ms/step - loss: 0.1145 - accuracy: 0.9663 - val_loss: 0.0054 - val_accuracy: 0.9985\n",
      "Epoch 15/20\n",
      "9583/9583 [==============================] - 92s 10ms/step - loss: 0.1025 - accuracy: 0.9686 - val_loss: 0.0066 - val_accuracy: 0.9983\n",
      "Epoch 16/20\n",
      "9583/9583 [==============================] - 90s 9ms/step - loss: 0.0927 - accuracy: 0.9713 - val_loss: 0.0036 - val_accuracy: 0.9993\n",
      "Epoch 17/20\n",
      "9583/9583 [==============================] - 90s 9ms/step - loss: 0.0880 - accuracy: 0.9723 - val_loss: 0.0054 - val_accuracy: 0.9988\n",
      "Epoch 18/20\n",
      "9583/9583 [==============================] - 101s 11ms/step - loss: 0.0864 - accuracy: 0.9729 - val_loss: 0.0042 - val_accuracy: 0.9988\n",
      "Epoch 19/20\n",
      "9583/9583 [==============================] - 94s 10ms/step - loss: 0.0839 - accuracy: 0.9750 - val_loss: 0.0040 - val_accuracy: 0.9988\n",
      "Epoch 20/20\n",
      "9583/9583 [==============================] - 93s 10ms/step - loss: 0.0817 - accuracy: 0.9745 - val_loss: 0.0038 - val_accuracy: 0.9988\n"
     ]
    }
   ],
   "source": [
    "checkpoint = ModelCheckpoint('model-{epoch:03d}.model',monitor='val_loss',verbose=0,save_best_only=True,mode='auto')\n",
    "history=classifier.fit(train_data,train_target,shuffle=True,epochs=20,callbacks=[checkpoint],validation_split=0.3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3423/3423 [==============================] - 8s 2ms/step\n",
      "[0.006450053481155843, 0.9985392689704895]\n"
     ]
    }
   ],
   "source": [
    "print(classifier.evaluate(test_data,test_target))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "N = 20\n",
    "H=history\n",
    "plt.style.use(\"ggplot\")\n",
    "plt.figure()\n",
    "plt.plot(np.arange(0, N), H.history[\"loss\"], label=\"train_loss\")\n",
    "plt.plot(np.arange(0, N), H.history[\"val_loss\"], label=\"val_loss\")\n",
    "plt.plot(np.arange(0, N), H.history[\"accuracy\"], label=\"train_acc\")\n",
    "plt.plot(np.arange(0, N), H.history[\"val_accuracy\"], label=\"val_acc\")\n",
    "plt.title(\"Training Loss and Accuracy\")\n",
    "plt.xlabel(\"Epoch #\")\n",
    "plt.ylabel(\"Loss/Accuracy\")\n",
    "plt.legend(loc=\"lower left\")\n",
    "plt.savefig('evaluation.png')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[INFO] saving mask detector model...\n",
      "Done !\n"
     ]
    }
   ],
   "source": [
    "# serialize the model to disk\n",
    "print(\"[INFO] saving mask detector model...\")\n",
    "classifier.save('asl_classifier.h5')\n",
    "print(\"Done !\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "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": [
    "import matplotlib.pyplot as plt\n",
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.xlabel('epochs')\n",
    "plt.ylabel('Loss')\n",
    "plt.legend(['train_loss','val_loss'], loc=0)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "plt.plot(history.history['accuracy'])\n",
    "plt.plot(history.history['val_accuracy'])\n",
    "plt.xlabel('epochs')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.legend(['train_accuracy','val_accuracy'], loc=0)\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": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}