Denoising.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>label</th>\n",
       "      <th>pixel0</th>\n",
       "      <th>pixel1</th>\n",
       "      <th>pixel2</th>\n",
       "      <th>pixel3</th>\n",
       "      <th>pixel4</th>\n",
       "      <th>pixel5</th>\n",
       "      <th>pixel6</th>\n",
       "      <th>pixel7</th>\n",
       "      <th>pixel8</th>\n",
       "      <th>...</th>\n",
       "      <th>pixel774</th>\n",
       "      <th>pixel775</th>\n",
       "      <th>pixel776</th>\n",
       "      <th>pixel777</th>\n",
       "      <th>pixel778</th>\n",
       "      <th>pixel779</th>\n",
       "      <th>pixel780</th>\n",
       "      <th>pixel781</th>\n",
       "      <th>pixel782</th>\n",
       "      <th>pixel783</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 785 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   label  pixel0  pixel1  pixel2  pixel3  pixel4  pixel5  pixel6  pixel7  \\\n",
       "0      1       0       0       0       0       0       0       0       0   \n",
       "1      0       0       0       0       0       0       0       0       0   \n",
       "2      1       0       0       0       0       0       0       0       0   \n",
       "3      4       0       0       0       0       0       0       0       0   \n",
       "4      0       0       0       0       0       0       0       0       0   \n",
       "\n",
       "   pixel8  ...  pixel774  pixel775  pixel776  pixel777  pixel778  pixel779  \\\n",
       "0       0  ...         0         0         0         0         0         0   \n",
       "1       0  ...         0         0         0         0         0         0   \n",
       "2       0  ...         0         0         0         0         0         0   \n",
       "3       0  ...         0         0         0         0         0         0   \n",
       "4       0  ...         0         0         0         0         0         0   \n",
       "\n",
       "   pixel780  pixel781  pixel782  pixel783  \n",
       "0         0         0         0         0  \n",
       "1         0         0         0         0  \n",
       "2         0         0         0         0  \n",
       "3         0         0         0         0  \n",
       "4         0         0         0         0  \n",
       "\n",
       "[5 rows x 785 columns]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = pd.read_csv(\"dataset/train.csv\")\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>pixel0</th>\n",
       "      <th>pixel1</th>\n",
       "      <th>pixel2</th>\n",
       "      <th>pixel3</th>\n",
       "      <th>pixel4</th>\n",
       "      <th>pixel5</th>\n",
       "      <th>pixel6</th>\n",
       "      <th>pixel7</th>\n",
       "      <th>pixel8</th>\n",
       "      <th>pixel9</th>\n",
       "      <th>...</th>\n",
       "      <th>pixel774</th>\n",
       "      <th>pixel775</th>\n",
       "      <th>pixel776</th>\n",
       "      <th>pixel777</th>\n",
       "      <th>pixel778</th>\n",
       "      <th>pixel779</th>\n",
       "      <th>pixel780</th>\n",
       "      <th>pixel781</th>\n",
       "      <th>pixel782</th>\n",
       "      <th>pixel783</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 784 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   pixel0  pixel1  pixel2  pixel3  pixel4  pixel5  pixel6  pixel7  pixel8  \\\n",
       "0       0       0       0       0       0       0       0       0       0   \n",
       "1       0       0       0       0       0       0       0       0       0   \n",
       "2       0       0       0       0       0       0       0       0       0   \n",
       "3       0       0       0       0       0       0       0       0       0   \n",
       "4       0       0       0       0       0       0       0       0       0   \n",
       "\n",
       "   pixel9  ...  pixel774  pixel775  pixel776  pixel777  pixel778  pixel779  \\\n",
       "0       0  ...         0         0         0         0         0         0   \n",
       "1       0  ...         0         0         0         0         0         0   \n",
       "2       0  ...         0         0         0         0         0         0   \n",
       "3       0  ...         0         0         0         0         0         0   \n",
       "4       0  ...         0         0         0         0         0         0   \n",
       "\n",
       "   pixel780  pixel781  pixel782  pixel783  \n",
       "0         0         0         0         0  \n",
       "1         0         0         0         0  \n",
       "2         0         0         0         0  \n",
       "3         0         0         0         0  \n",
       "4         0         0         0         0  \n",
       "\n",
       "[5 rows x 784 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "t_data = data.drop([\"label\"],axis = 1)\n",
    "t_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "t_data = np.matrix(t_data)/255\n",
    "mu, sigma = 0, 0.1 \n",
    "noise = np.random.normal(mu, sigma, [len(t_data),784])\n",
    "n_data = t_data + noise"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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": [
    "tmp = t_data[1]\n",
    "tmp.shape = (28,28)\n",
    "plt.imshow(255-tmp,cmap='gray')\n",
    "plt.xlabel('original')\n",
    "plt.show()\n",
    "\n",
    "tmp = n_data[1]\n",
    "tmp.shape = (28,28)\n",
    "plt.imshow(255-tmp,cmap='gray')\n",
    "plt.xlabel('noisy')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = n_data[:30000]\n",
    "y_train = t_data[:30000]\n",
    "X_test = n_data[30000:]\n",
    "y_test = t_data[30000:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 27000 samples, validate on 3000 samples\n",
      "Epoch 1/10\n",
      "27000/27000 [==============================] - 16s 609us/sample - loss: 0.1345 - acc: 0.8073 - val_loss: 0.0963 - val_acc: 0.8123\n",
      "Epoch 2/10\n",
      "27000/27000 [==============================] - 16s 585us/sample - loss: 0.0891 - acc: 0.8142 - val_loss: 0.0858 - val_acc: 0.8131\n",
      "Epoch 3/10\n",
      "27000/27000 [==============================] - 16s 588us/sample - loss: 0.0827 - acc: 0.8147 - val_loss: 0.0817 - val_acc: 0.8133\n",
      "Epoch 4/10\n",
      "27000/27000 [==============================] - 16s 588us/sample - loss: 0.0795 - acc: 0.8149 - val_loss: 0.0796 - val_acc: 0.8134\n",
      "Epoch 5/10\n",
      "27000/27000 [==============================] - 16s 599us/sample - loss: 0.0775 - acc: 0.8150 - val_loss: 0.0780 - val_acc: 0.8135\n",
      "Epoch 6/10\n",
      "27000/27000 [==============================] - 17s 632us/sample - loss: 0.0762 - acc: 0.8150 - val_loss: 0.0775 - val_acc: 0.8135\n",
      "Epoch 7/10\n",
      "27000/27000 [==============================] - 16s 608us/sample - loss: 0.0753 - acc: 0.8151 - val_loss: 0.0768 - val_acc: 0.8136\n",
      "Epoch 8/10\n",
      "27000/27000 [==============================] - 16s 591us/sample - loss: 0.0745 - acc: 0.8151 - val_loss: 0.0756 - val_acc: 0.8136\n",
      "Epoch 9/10\n",
      "27000/27000 [==============================] - 18s 653us/sample - loss: 0.0740 - acc: 0.8151 - val_loss: 0.0754 - val_acc: 0.8136\n",
      "Epoch 10/10\n",
      "27000/27000 [==============================] - 16s 607us/sample - loss: 0.0735 - acc: 0.8151 - val_loss: 0.0753 - val_acc: 0.8136\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x7f770410f1d0>"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense, Dropout, Activation, Flatten\n",
    "\n",
    "model = Sequential()\n",
    "model.add(Flatten())\n",
    "model.add(Dense(300))\n",
    "model.add(Activation(\"relu\"))\n",
    "model.add(Dense(150))\n",
    "model.add(Activation(\"relu\"))\n",
    "model.add(Dense(300))\n",
    "model.add(Activation(\"relu\"))\n",
    "model.add(Dense(784))\n",
    "model.add(Activation(\"sigmoid\"))\n",
    "#model.add(Dropout(0.2))\n",
    "model.compile(loss=\"binary_crossentropy\", optimizer=\"adam\", metrics=['accuracy'])\n",
    "model.fit(X_train, y_train, batch_size=32, epochs=10, validation_split=0.1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "12000/12000 [==============================] - 2s 158us/sample - loss: 0.0749 - acc: 0.8154\n"
     ]
    }
   ],
   "source": [
    "val_loss, val_accuracy = model.evaluate(X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "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"
    },
    {
     "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": [
    "tmp = y_test[5]\n",
    "tmp.shape = (28,28)\n",
    "plt.imshow(255-tmp,cmap='gray')\n",
    "plt.xlabel('Original')\n",
    "plt.show()\n",
    "\n",
    "tmp = X_test[5]\n",
    "tmp.shape = (28,28)\n",
    "plt.imshow(255-tmp,cmap='gray')\n",
    "plt.xlabel('Noisy')\n",
    "plt.show()\n",
    "\n",
    "prediction = model.predict([[tmp]])\n",
    "prediction.shape = (28,28)\n",
    "plt.imshow(255-prediction,cmap='gray')\n",
    "plt.xlabel('After Denoising')\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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}