material/deep_learning/mnist_super_resolution.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "mnist_super_resolution.ipynb",
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jLKCaOxfGQg3"
      },
      "source": [
        "modified from https://blog.naver.com/ckdgus1433/221443838135"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PhlLzK2qzOQU"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "from tensorflow.keras.datasets import mnist\n",
        "from tensorflow.keras.layers import Dense\n",
        "from tensorflow.keras import Sequential"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HHW1dtOvzOTu",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "fe816bd7-5509-49d7-c996-51c720eed93e"
      },
      "source": [
        "# MNIST 로딩 (라벨은 필요없기 때문에 버림)\n",
        "(train_x, _), (test_x, _) = mnist.load_data()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/mnist.npz\n",
            "11493376/11490434 [==============================] - 0s 0us/step\n",
            "11501568/11490434 [==============================] - 0s 0us/step\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "E5B1_T2yzONm"
      },
      "source": [
        "# 데이터 정규화 및 Reshape\n",
        "train_x = train_x.astype('float32') / 255.\n",
        "test_x = test_x.astype('float32') / 255."
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9I4G_rT-zecq",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "1351319a-0991-451e-8188-a98cd5b8f7ea"
      },
      "source": [
        "def shrink(data, rows, cols):\n",
        "    return data.reshape(data.shape[0], rows, int(data.shape[1]/rows), cols, int(data.shape[2]/cols)).sum(axis=2).sum(axis=3)\n",
        "\n",
        "reduced_train_x = shrink(train_x, 14, 14)\n",
        "reduced_test_x = shrink(test_x, 14, 14)\n",
        "print(reduced_train_x.shape)\n",
        "print(reduced_test_x.shape)\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(60000, 14, 14)\n",
            "(10000, 14, 14)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5jFitrsf13K3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 241
        },
        "outputId": "8d370583-2267-46f2-89f9-14cd6f4f8f9d"
      },
      "source": [
        "n = 10\n",
        "\n",
        "plt.figure(figsize=(20, 2))\n",
        "for i in range(1,n):\n",
        "    ax = plt.subplot(1, n, i)\n",
        "    plt.imshow(train_x[i])\n",
        "    plt.gray()\n",
        "    ax.get_xaxis().set_visible(False)\n",
        "    ax.get_yaxis().set_visible(False)\n",
        "plt.show()\n",
        "\n",
        "# 해상도 낮아진 이미지 확인\n",
        "plt.figure(figsize=(20, 2))\n",
        "for i in range(1,n):\n",
        "    ax = plt.subplot(1, n, i)\n",
        "    plt.imshow(reduced_train_x[i])\n",
        "    plt.gray()\n",
        "    ax.get_xaxis().set_visible(False)\n",
        "    ax.get_yaxis().set_visible(False)\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1440x144 with 9 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1440x144 with 9 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QtW_31_Q2IG6"
      },
      "source": [
        "reduced_train_x = reduced_train_x.reshape(reduced_train_x.shape[0], 14*14)\n",
        "reduced_test_x = reduced_test_x.reshape(reduced_test_x.shape[0], 14*14)\n",
        "train_x = train_x.reshape(train_x.shape[0], 28*28)\n",
        "test_x = test_x.reshape(test_x.shape[0], 28*28)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gUTop4Mg2z0T",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d783d8cf-db72-499d-eead-bbf7fdc28d2f"
      },
      "source": [
        "print(reduced_train_x.shape)\n",
        "print(train_x.shape)\n",
        "print(reduced_test_x.shape)\n",
        "print(test_x.shape)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(60000, 196)\n",
            "(60000, 784)\n",
            "(10000, 196)\n",
            "(10000, 784)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Xqt71jtz2fjo",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "20b74839-0c5e-47f3-8006-27a3c7129c5f"
      },
      "source": [
        "# 모델 정의\n",
        "model = Sequential()\n",
        "\n",
        "model.add(Dense(128, activation='relu', input_dim=14*14))\n",
        "model.add(Dense(64, activation='relu'))\n",
        "model.add(Dense(32, activation='relu'))\n",
        "model.add(Dense(64, activation='relu'))\n",
        "model.add(Dense(128, activation='relu'))\n",
        "model.add(Dense(784, activation='sigmoid'))\n",
        "\n",
        "model.compile(optimizer='adam', loss='mse')\n",
        "\n",
        "# 모델 학습\n",
        "model.fit(reduced_train_x, train_x, epochs=50, batch_size=256, validation_split=0.1)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/50\n",
            "211/211 [==============================] - 1s 4ms/step - loss: 0.0640 - val_loss: 0.0370\n",
            "Epoch 2/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0322 - val_loss: 0.0283\n",
            "Epoch 3/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0258 - val_loss: 0.0231\n",
            "Epoch 4/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0214 - val_loss: 0.0196\n",
            "Epoch 5/50\n",
            "211/211 [==============================] - 1s 4ms/step - loss: 0.0189 - val_loss: 0.0178\n",
            "Epoch 6/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0172 - val_loss: 0.0163\n",
            "Epoch 7/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0161 - val_loss: 0.0155\n",
            "Epoch 8/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0152 - val_loss: 0.0145\n",
            "Epoch 9/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0144 - val_loss: 0.0140\n",
            "Epoch 10/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0139 - val_loss: 0.0135\n",
            "Epoch 11/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0134 - val_loss: 0.0132\n",
            "Epoch 12/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0130 - val_loss: 0.0128\n",
            "Epoch 13/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0127 - val_loss: 0.0124\n",
            "Epoch 14/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0123 - val_loss: 0.0121\n",
            "Epoch 15/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0120 - val_loss: 0.0118\n",
            "Epoch 16/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0117 - val_loss: 0.0116\n",
            "Epoch 17/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0114 - val_loss: 0.0114\n",
            "Epoch 18/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0112 - val_loss: 0.0112\n",
            "Epoch 19/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0110 - val_loss: 0.0111\n",
            "Epoch 20/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0109 - val_loss: 0.0110\n",
            "Epoch 21/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0107 - val_loss: 0.0108\n",
            "Epoch 22/50\n",
            "211/211 [==============================] - 1s 4ms/step - loss: 0.0106 - val_loss: 0.0107\n",
            "Epoch 23/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0105 - val_loss: 0.0105\n",
            "Epoch 24/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0103 - val_loss: 0.0104\n",
            "Epoch 25/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0102 - val_loss: 0.0102\n",
            "Epoch 26/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0101 - val_loss: 0.0102\n",
            "Epoch 27/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0100 - val_loss: 0.0100\n",
            "Epoch 28/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0099 - val_loss: 0.0100\n",
            "Epoch 29/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0098 - val_loss: 0.0099\n",
            "Epoch 30/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0097 - val_loss: 0.0098\n",
            "Epoch 31/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0096 - val_loss: 0.0097\n",
            "Epoch 32/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0096 - val_loss: 0.0096\n",
            "Epoch 33/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0095 - val_loss: 0.0095\n",
            "Epoch 34/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0094 - val_loss: 0.0096\n",
            "Epoch 35/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0093 - val_loss: 0.0094\n",
            "Epoch 36/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0092 - val_loss: 0.0093\n",
            "Epoch 37/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0091 - val_loss: 0.0091\n",
            "Epoch 38/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0090 - val_loss: 0.0092\n",
            "Epoch 39/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0089 - val_loss: 0.0090\n",
            "Epoch 40/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0089 - val_loss: 0.0090\n",
            "Epoch 41/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0088 - val_loss: 0.0090\n",
            "Epoch 42/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0088 - val_loss: 0.0088\n",
            "Epoch 43/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0087 - val_loss: 0.0088\n",
            "Epoch 44/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0087 - val_loss: 0.0088\n",
            "Epoch 45/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0086 - val_loss: 0.0087\n",
            "Epoch 46/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0086 - val_loss: 0.0087\n",
            "Epoch 47/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0085 - val_loss: 0.0087\n",
            "Epoch 48/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0085 - val_loss: 0.0087\n",
            "Epoch 49/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0085 - val_loss: 0.0087\n",
            "Epoch 50/50\n",
            "211/211 [==============================] - 1s 3ms/step - loss: 0.0084 - val_loss: 0.0085\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7f7ebc736190>"
            ]
          },
          "metadata": {},
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2y4hQ-VfFxF4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 359
        },
        "outputId": "0c0371e0-70bc-4738-f834-7e57b0a366de"
      },
      "source": [
        "# 결과 확인\n",
        "super_resolutioned = model.predict(reduced_test_x)\n",
        "\n",
        "n = 10\n",
        "plt.figure(figsize=(20, 6))\n",
        "for i in range(1, n):\n",
        "    # display original\n",
        "    ax = plt.subplot(3, n, i)\n",
        "    plt.imshow(test_x[i].reshape(28,28))\n",
        "    plt.gray()\n",
        "    ax.get_xaxis().set_visible(False)\n",
        "    ax.get_yaxis().set_visible(False)\n",
        "\n",
        "    # display reduced\n",
        "    ax = plt.subplot(3, n, i + n)\n",
        "    plt.imshow(reduced_test_x[i].reshape(14,14))\n",
        "    plt.gray()\n",
        "    ax.get_xaxis().set_visible(False)\n",
        "    ax.get_yaxis().set_visible(False)\n",
        "\n",
        "    # display super resolution\n",
        "    ax = plt.subplot(3, n, i + 2*n)\n",
        "    plt.imshow(super_resolutioned[i].reshape(28,28))\n",
        "    plt.gray()\n",
        "    ax.get_xaxis().set_visible(False)\n",
        "    ax.get_yaxis().set_visible(False)\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1440x432 with 27 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fTqVK4F6F1S2"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}