auto_encoder.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "auto-encoder.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/kamzon/auto-encoder/blob/main/auto_encoder.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QYfqaxRbwOHT"
      },
      "source": [
        "# libraries\n",
        "import numpy as np\n",
        "from matplotlib import pyplot as plt \n",
        "import math\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "V8zrawAOxBIc"
      },
      "source": [
        "from keras.datasets import mnist\n",
        "(train_images,train_labels),(test_images,test_labels)= mnist.load_data()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fZUCESb1xcIc",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "0e179a68-bb4b-4327-8410-f1624fe74378"
      },
      "source": [
        "train_images.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(60000, 28, 28)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XW-1I6V_yf79"
      },
      "source": [
        "\n",
        "image_size=train_images.shape[1]\n",
        "train_images=np.reshape(train_images,[-1,image_size,image_size,1])\n",
        "test_images=np.reshape(test_images,[-1,image_size,image_size,1])\n",
        "\n",
        "train_images=train_images.astype('float32')/255\n",
        "test_images=test_images.astype('float32')/255\n",
        "\n",
        "#image avec noise\n",
        "\n",
        "noise=np.random.normal(loc=0, scale=0.1, size=train_images.shape)\n",
        "train_images_noisy=train_images + noise/10\n",
        "\n",
        "noise=np.random.normal(loc=0, scale=0.1, size=test_images.shape)\n",
        "test_images_noisy=test_images + noise/10\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kNp_nFSYlkF7"
      },
      "source": [
        "noise"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "N7mdP5jV7Xyu",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "outputId": "fb0543d9-eb73-4611-8989-4d43467cb0d5"
      },
      "source": [
        "plt.imshow(test_images_noisy[10].reshape(28,28),cmap='gray')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7fa6818bd978>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 210
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAPsAAAD4CAYAAAAq5pAIAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQskukgklu0gm/h9EzbuBqGErZQAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gkLy__fg9sWZ"
      },
      "source": [
        "import keras\n",
        "from keras.models import Sequential,Input,Model\n",
        "from keras.layers import Dense,Dropout,Flatten\n",
        "from keras.layers import Conv2D,MaxPooling2D\n",
        "from keras.layers.normalization import BatchNormalization\n",
        "from keras.layers.advanced_activations import LeakyReLU\n",
        "import keras.layers\n",
        "from keras.layers import Conv2DTranspose, Reshape\n",
        "from keras.layers import Activation, Dense\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0VlmM24E7XuH"
      },
      "source": [
        "\n",
        "\n",
        "#encoder:\n",
        "autoencoder_model=Sequential()\n",
        "autoencoder_model.add(Conv2D(64,kernel_size=5,strides=2,activation='relu',input_shape=(28,28,1),padding='same'))\n",
        "autoencoder_model.add(Conv2D(32,kernel_size=5,strides=2, activation='relu',padding='same'))\n",
        "autoencoder_model.add(Flatten())\n",
        "autoencoder_model.add(Dense(16))\n",
        "#autoencoder_model.add(Conv2D(64,kernel_size=5,strides=2,activation='relu',padding='same'))\n",
        "\n",
        "\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4N-0gR8w9rra"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1MT3u7Bb9ODX"
      },
      "source": [
        "#decoder\n",
        "autoencoder_model.add(Dense(1568))\n",
        "autoencoder_model.add(Reshape((7,7,32)))\n",
        "autoencoder_model.add(Conv2DTranspose(32,kernel_size=5,strides=2,activation='relu',padding='same'))\n",
        "\n",
        "autoencoder_model.add(Conv2DTranspose(64,kernel_size=5,strides=2,activation='relu',padding='same'))\n",
        "\n",
        "#autoencoder_model.add(Conv2DTranspose(16,kernel_size=5,strides=2,activation='relu',padding='same'))\n",
        "\n",
        "autoencoder_model.add(Conv2DTranspose(1,kernel_size=5,padding='same'))\n",
        "\n",
        "autoencoder_model.add(Activation('sigmoid'))\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sdf3eWbDBsct"
      },
      "source": [
        "autoencoder_model.compile(loss='mse',optimizer='adam')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "URp__ux7CC0f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 493
        },
        "outputId": "ac4835b4-9676-4318-ebac-9b5d64230b82"
      },
      "source": [
        "autoencoder_model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_25\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv2d_49 (Conv2D)           (None, 14, 14, 64)        1664      \n",
            "_________________________________________________________________\n",
            "conv2d_50 (Conv2D)           (None, 7, 7, 32)          51232     \n",
            "_________________________________________________________________\n",
            "flatten_10 (Flatten)         (None, 1568)              0         \n",
            "_________________________________________________________________\n",
            "dense_16 (Dense)             (None, 16)                25104     \n",
            "_________________________________________________________________\n",
            "dense_17 (Dense)             (None, 1568)              26656     \n",
            "_________________________________________________________________\n",
            "reshape_5 (Reshape)          (None, 7, 7, 32)          0         \n",
            "_________________________________________________________________\n",
            "conv2d_transpose_66 (Conv2DT (None, 14, 14, 32)        25632     \n",
            "_________________________________________________________________\n",
            "conv2d_transpose_67 (Conv2DT (None, 28, 28, 64)        51264     \n",
            "_________________________________________________________________\n",
            "conv2d_transpose_68 (Conv2DT (None, 28, 28, 1)         1601      \n",
            "_________________________________________________________________\n",
            "activation_10 (Activation)   (None, 28, 28, 1)         0         \n",
            "=================================================================\n",
            "Total params: 183,153\n",
            "Trainable params: 183,153\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7Bt-JLTZCNVV",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 391
        },
        "outputId": "c66f0a9d-b5de-41d5-e2c9-db73f8e3445f"
      },
      "source": [
        "autoencoder_train=autoencoder_model.fit(train_images_noisy,train_images, batch_size=32,nb_epoch=10,verbose=1)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:1: UserWarning: The `nb_epoch` argument in `fit` has been renamed `epochs`.\n",
            "  \"\"\"Entry point for launching an IPython kernel.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/10\n",
            "60000/60000 [==============================] - 15s 245us/step - loss: 0.0197\n",
            "Epoch 2/10\n",
            "60000/60000 [==============================] - 13s 209us/step - loss: 0.0096\n",
            "Epoch 3/10\n",
            "60000/60000 [==============================] - 13s 215us/step - loss: 0.0085\n",
            "Epoch 4/10\n",
            "60000/60000 [==============================] - 13s 211us/step - loss: 0.0079\n",
            "Epoch 5/10\n",
            "60000/60000 [==============================] - 13s 211us/step - loss: 0.0075\n",
            "Epoch 6/10\n",
            "60000/60000 [==============================] - 13s 211us/step - loss: 0.0072\n",
            "Epoch 7/10\n",
            "60000/60000 [==============================] - 13s 213us/step - loss: 0.0070\n",
            "Epoch 8/10\n",
            "60000/60000 [==============================] - 13s 215us/step - loss: 0.0068\n",
            "Epoch 9/10\n",
            "60000/60000 [==============================] - 13s 217us/step - loss: 0.0067\n",
            "Epoch 10/10\n",
            "60000/60000 [==============================] - 13s 212us/step - loss: 0.0066\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VEhJl825h-Do",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 147
        },
        "outputId": "cc685476-1ddb-45d1-a861-ae74327c6034"
      },
      "source": [
        "from PIL import Image\n",
        "\n",
        "\n",
        "# Prediction the Autoencoder output -- test images\n",
        "x_decoded = autoencoder_model.predict(test_images_noisy)\n",
        "\n",
        "image_size = train_images.shape[1]\n",
        "rows, cols = 1, 10\n",
        "num = rows * cols\n",
        "imgs = np.concatenate([test_images[:num], test_images_noisy[:num], x_decoded[:num]])\n",
        "imgs = imgs.reshape((rows * 3, cols, image_size, image_size))\n",
        "imgs = np.vstack(np.split(imgs, rows, axis=1))\n",
        "imgs = imgs.reshape((rows * 3, -1, image_size, image_size))\n",
        "imgs = np.vstack([np.hstack(i) for i in imgs])\n",
        "imgs = (imgs * 255).astype(np.uint8)\n",
        "plt.figure()\n",
        "plt.axis('off')\n",
        "plt.title('Original images: top rows, '\n",
        "          'Corrupted Input: middle rows, '\n",
        "          'Denoised Input:  third rows')\n",
        "plt.imshow(imgs, interpolation='none', cmap='gray')\n",
        "Image.fromarray(imgs).save('corrupted_and_denoised.png')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAg0AAACCCAYAAADFYfuwAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cybts5hs2Md-",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "outputId": "69eebf51-a0ac-4b32-9018-78bce0f46e55"
      },
      "source": [
        "plt.imshow(test_images[100].reshape(28,28))\n",
        "\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7fa6819cf9e8>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 203
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAPsAAAD4CAYAAAAq5pAIAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "25-IvK-S2eGA",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "outputId": "f35ff3aa-7d01-4b1d-df1d-379d3fbb6b32"
      },
      "source": [
        "plt.imshow(test_images_noisy[100].reshape(28,28))\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7fa6819ad5f8>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 204
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAPsAAAD4CAYAAAAq5pAIAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BJhXYkJE2gc2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "outputId": "248d31e6-3337-4f5d-a95f-1f7f0f7e4c5c"
      },
      "source": [
        "plt.imshow(x_decoded[100].reshape(28,28))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7fa6819886d8>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 205
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAPsAAAD4CAYAAAAq5pAIAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nA0kQdiAJwg4k8f9oPb+RQbF9OAAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    }
  ]
}