MedicineTransferLearning.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "id": "MP_24u09U7hQ",
    "colab_type": "code",
    "colab": {}
   },
   "outputs": [],
   "source": [
    "from keras.applications.resnet50 import ResNet50\n",
    "from keras.preprocessing import image\n",
    "from keras.models import Model\n",
    "from keras.layers import Dense, GlobalAveragePooling2D\n",
    "from keras import backend as K\n",
    "from keras.applications.resnet50 import preprocess_input, decode_predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "id": "QmZd84uYVcE_",
    "colab_type": "code",
    "outputId": "5c962eb9-c9bb-4ba4-f6f4-d3df58314dee",
    "executionInfo": {
     "status": "ok",
     "timestamp": 1.555300742983E12,
     "user_tz": -480.0,
     "elapsed": 14448.0,
     "user": {
      "displayName": "J.L H",
      "photoUrl": "https://lh5.googleusercontent.com/-BRX5TWuOHpE/AAAAAAAAAAI/AAAAAAAAAD0/SzepdPSA1LY/s64/photo.jpg",
      "userId": "14158868708361974289"
     }
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 105.0
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.6/dist-packages/keras_applications/resnet50.py:265: UserWarning: The output shape of `ResNet50(include_top=False)` has been changed since Keras 2.2.0.\n",
      "  warnings.warn('The output shape of `ResNet50(include_top=False)` '\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading data from https://github.com/fchollet/deep-learning-models/releases/download/v0.2/resnet50_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
      "94658560/94653016 [==============================] - 1s 0us/step\n"
     ]
    }
   ],
   "source": [
    "# 构建不带分类器的预训练模型\n",
    "base_model = ResNet50(weights='imagenet', include_top=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "id": "4K48bKYfV6o6",
    "colab_type": "code",
    "outputId": "6ac0be65-4fb4-47cc-9a97-8175f5760e55",
    "executionInfo": {
     "status": "ok",
     "timestamp": 1.555300751485E12,
     "user_tz": -480.0,
     "elapsed": 5501.0,
     "user": {
      "displayName": "J.L H",
      "photoUrl": "https://lh5.googleusercontent.com/-BRX5TWuOHpE/AAAAAAAAAAI/AAAAAAAAAD0/SzepdPSA1LY/s64/photo.jpg",
      "userId": "14158868708361974289"
     }
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34.0
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "resnet50_weights_tf_dim_ordering_tf_kernels_notop.h5\n"
     ]
    }
   ],
   "source": [
    "!ls ~/.keras/models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "id": "y3Oz5X9HWCzf",
    "colab_type": "code",
    "outputId": "d7605d4e-274e-4d31-ee9e-0fbcddac8263",
    "executionInfo": {
     "status": "ok",
     "timestamp": 1.555300646222E12,
     "user_tz": -480.0,
     "elapsed": 27263.0,
     "user": {
      "displayName": "J.L H",
      "photoUrl": "https://lh5.googleusercontent.com/-BRX5TWuOHpE/AAAAAAAAAAI/AAAAAAAAAD0/SzepdPSA1LY/s64/photo.jpg",
      "userId": "14158868708361974289"
     }
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 122.0
    },
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Go to this URL in a browser: https://accounts.google.com/o/oauth2/auth?client_id=947318989803-6bn6qk8qdgf4n4g3pfee6491hc0brc4i.apps.googleusercontent.com&redirect_uri=urn%3Aietf%3Awg%3Aoauth%3A2.0%3Aoob&scope=email%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fdocs.test%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fdrive%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fdrive.photos.readonly%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fpeopleapi.readonly&response_type=code\n",
      "\n",
      "Enter your authorization code:\n",
      "··········\n",
      "Mounted at /content/gdrive\n"
     ]
    }
   ],
   "source": [
    "from google.colab import drive\n",
    "drive.mount('/content/gdrive', force_remount=True)\n",
    "root_dir = \"/content/gdrive/My Drive/\"\n",
    "base_dir = root_dir + 'data/medicine/'\n",
    "TRAIN_DIR = base_dir + 'train/'\n",
    "VAL_DIR = base_dir + 'test/'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "id": "qxsazmdkWPxI",
    "colab_type": "code",
    "outputId": "8d03978c-7f20-4499-8640-1f68a0c192fa",
    "executionInfo": {
     "status": "ok",
     "timestamp": 1.555300759282E12,
     "user_tz": -480.0,
     "elapsed": 2291.0,
     "user": {
      "displayName": "J.L H",
      "photoUrl": "https://lh5.googleusercontent.com/-BRX5TWuOHpE/AAAAAAAAAAI/AAAAAAAAAD0/SzepdPSA1LY/s64/photo.jpg",
      "userId": "14158868708361974289"
     }
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34.0
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 362 images belonging to 3 classes.\n"
     ]
    }
   ],
   "source": [
    "IMG_W,IMG_H =224,224 # 单张图片的大小\n",
    "BATCH_SIZE = 16\n",
    "# 2，准备训练集，keras有很多Generator可以直接处理图片的加载，增强等操作，封装的非常好\n",
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "train_datagen = ImageDataGenerator( # 单张图片的处理方式，train时一般都会进行图片增强\n",
    "        rescale=1. / 255, # 图片像素值为0-255，此处都乘以1/255，调整到0-1之间\n",
    "        shear_range=0.2, # 斜切\n",
    "        zoom_range=0.2, # 放大缩小范围\n",
    "        horizontal_flip=True) # 水平翻转\n",
    "\n",
    "train_generator = train_datagen.flow_from_directory(# 从文件夹中产生数据流\n",
    "    TRAIN_DIR, # 训练集图片的文件夹\n",
    "    target_size=(IMG_W, IMG_H), # 调整后每张图片的大小\n",
    "    batch_size=BATCH_SIZE,\n",
    "    class_mode='categorical') # 此处是多分类问题，故而mode是categorical\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "id": "dYI4Gf47W65F",
    "colab_type": "code",
    "outputId": "b7376068-fb47-42db-a725-98bbfa26ca87",
    "executionInfo": {
     "status": "ok",
     "timestamp": 1.555300763573E12,
     "user_tz": -480.0,
     "elapsed": 1135.0,
     "user": {
      "displayName": "J.L H",
      "photoUrl": "https://lh5.googleusercontent.com/-BRX5TWuOHpE/AAAAAAAAAAI/AAAAAAAAAD0/SzepdPSA1LY/s64/photo.jpg",
      "userId": "14158868708361974289"
     }
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34.0
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 84 images belonging to 3 classes.\n"
     ]
    }
   ],
   "source": [
    "# 3，同样的方式准备测试集\n",
    "val_datagen = ImageDataGenerator(rescale=1. / 255) # 只需要和trainset同样的scale即可，不需增强\n",
    "val_generator = val_datagen.flow_from_directory(\n",
    "        VAL_DIR,\n",
    "        target_size=(IMG_W, IMG_H),\n",
    "        batch_size=BATCH_SIZE,\n",
    "        class_mode='categorical')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "id": "z_6vq-MSYdsJ",
    "colab_type": "code",
    "outputId": "d06804ea-6147-4671-b663-9d6f794269ab",
    "executionInfo": {
     "status": "ok",
     "timestamp": 1.555300770942E12,
     "user_tz": -480.0,
     "elapsed": 1035.0,
     "user": {
      "displayName": "J.L H",
      "photoUrl": "https://lh5.googleusercontent.com/-BRX5TWuOHpE/AAAAAAAAAAI/AAAAAAAAAD0/SzepdPSA1LY/s64/photo.jpg",
      "userId": "14158868708361974289"
     }
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 51.0
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3\n",
      "{'枸杞子': 0, '白术 ': 1, '茯苓': 2}\n"
     ]
    }
   ],
   "source": [
    "#统计分类数\n",
    "num_classes = len(train_generator.class_indices)\n",
    "class_dictionary = train_generator.class_indices\n",
    "print(num_classes)\n",
    "print(class_dictionary)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "id": "s6UpPcIjYheB",
    "colab_type": "code",
    "colab": {}
   },
   "outputs": [],
   "source": [
    "from keras import optimizers\n",
    "# 添加全局平均池化层\n",
    "x = base_model.output\n",
    "x = GlobalAveragePooling2D()(x)\n",
    "\n",
    "# 添加一个全连接层\n",
    "x = Dense(512, activation='relu')(x)\n",
    "\n",
    "# 添加一个分类器，假设我们有3个类\n",
    "predictions = Dense(3, activation='softmax')(x)\n",
    "\n",
    "# 构建我们需要训练的完整模型\n",
    "model = Model(inputs=base_model.input, outputs=predictions)\n",
    "\n",
    "# 首先，我们只训练顶部的几层（随机初始化的层）\n",
    "# 锁住所有 InceptionV3 的卷积层\n",
    "# for layer in base_model.layers:\n",
    "#     layer.trainable = False\n",
    "\n",
    "# 编译模型（一定要在锁层以后操作）\n",
    "model.compile(loss='categorical_crossentropy',\n",
    "              optimizer=optimizers.SGD(lr=1e-4, momentum=0.9),\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "id": "Bb3ekHpCaZka",
    "colab_type": "code",
    "outputId": "75697658-4fbb-4788-9a69-6a80d655b8ad",
    "executionInfo": {
     "status": "ok",
     "timestamp": 1.555255365303E12,
     "user_tz": -480.0,
     "elapsed": 119683.0,
     "user": {
      "displayName": "J.L H",
      "photoUrl": "https://lh5.googleusercontent.com/-BRX5TWuOHpE/AAAAAAAAAAI/AAAAAAAAAD0/SzepdPSA1LY/s64/photo.jpg",
      "userId": "14158868708361974289"
     }
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 411.0
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.6/dist-packages/PIL/Image.py:914: UserWarning: Palette images with Transparency   expressed in bytes should be converted to RGBA images\n",
      "  'to RGBA images')\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "22/22 [==============================] - 12s 552ms/step - loss: 0.2469 - acc: 0.9643 - val_loss: 0.2140 - val_acc: 0.9265\n",
      "Epoch 2/10\n",
      "22/22 [==============================] - 12s 531ms/step - loss: 0.1807 - acc: 0.9813 - val_loss: 0.1931 - val_acc: 0.9706\n",
      "Epoch 3/10\n",
      "22/22 [==============================] - 12s 534ms/step - loss: 0.1827 - acc: 0.9744 - val_loss: 0.1512 - val_acc: 0.9559\n",
      "Epoch 4/10\n",
      "22/22 [==============================] - 12s 529ms/step - loss: 0.1506 - acc: 0.9829 - val_loss: 0.1515 - val_acc: 0.9706\n",
      "Epoch 5/10\n",
      "22/22 [==============================] - 12s 553ms/step - loss: 0.1269 - acc: 0.9886 - val_loss: 0.1375 - val_acc: 0.9750\n",
      "Epoch 6/10\n",
      "22/22 [==============================] - 12s 526ms/step - loss: 0.1260 - acc: 0.9825 - val_loss: 0.1171 - val_acc: 0.9853\n",
      "Epoch 7/10\n",
      "22/22 [==============================] - 12s 543ms/step - loss: 0.1058 - acc: 0.9886 - val_loss: 0.1263 - val_acc: 0.9559\n",
      "Epoch 8/10\n",
      "22/22 [==============================] - 12s 551ms/step - loss: 0.0951 - acc: 0.9825 - val_loss: 0.0899 - val_acc: 0.9853\n",
      "Epoch 9/10\n",
      "22/22 [==============================] - 11s 518ms/step - loss: 0.0837 - acc: 0.9886 - val_loss: 0.0936 - val_acc: 0.9853\n",
      "Epoch 10/10\n",
      "22/22 [==============================] - 12s 538ms/step - loss: 0.1098 - acc: 0.9773 - val_loss: 0.1112 - val_acc: 0.9706\n"
     ]
    }
   ],
   "source": [
    "# 模型的训练\n",
    "NUM_EPOCHS = 10\n",
    "BATCH_SIZE = 16\n",
    "num_train_images = 362\n",
    "num_val_images = 84\n",
    "history_1 = model.fit_generator(train_generator, # 数据流\n",
    "                        steps_per_epoch=num_train_images // BATCH_SIZE, \n",
    "                        epochs=NUM_EPOCHS,\n",
    "                        shuffle = True, \n",
    "                        verbose = 1, \n",
    "                        validation_data=val_generator,\n",
    "                        validation_steps=num_val_images // BATCH_SIZE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "id": "Da4HzFercfAd",
    "colab_type": "code",
    "colab": {}
   },
   "outputs": [],
   "source": [
    "# 画图，将训练时的acc和loss都绘制到图上\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "%matplotlib inline\n",
    "\n",
    "def plot_training(history):\n",
    "    plt.figure(figsize=(8, 8))\n",
    "    plt.subplot(2, 1, 1)\n",
    "    train_acc = history.history['acc']\n",
    "    val_acc = history.history['val_acc']\n",
    "    #epochs = range(len(train_acc))\n",
    "    plt.plot(train_acc,label='Training Accuracy')\n",
    "    plt.plot(val_acc,label='Validation Accuracy')\n",
    "    plt.legend(loc='lower right')\n",
    "    plt.ylabel('Accuracy')\n",
    "    plt.xlabel('epoch')\n",
    "    plt.title('Training and Validation Accuracy')\n",
    "    \n",
    "    plt.subplot(2, 1, 2)\n",
    "    train_loss = history.history['loss']\n",
    "    val_loss = history.history['val_loss']\n",
    "    #epochs = range(len(train_loss))\n",
    "    plt.plot(train_loss, label='Training Loss')\n",
    "    plt.plot(val_loss, label='Validation Loss')\n",
    "    plt.legend(loc='upper right')\n",
    "    plt.ylabel('Cross Entropy')\n",
    "    plt.title('Training and Validation Loss')\n",
    "    plt.xlabel('epoch')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "id": "Yb1zeXWbcgCy",
    "colab_type": "code",
    "outputId": "e65b8750-9839-4fee-9384-7f013cf52f98",
    "executionInfo": {
     "status": "ok",
     "timestamp": 1.555255429068E12,
     "user_tz": -480.0,
     "elapsed": 2253.0,
     "user": {
      "displayName": "J.L H",
      "photoUrl": "https://lh5.googleusercontent.com/-BRX5TWuOHpE/AAAAAAAAAAI/AAAAAAAAAD0/SzepdPSA1LY/s64/photo.jpg",
      "userId": "14158868708361974289"
     }
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 513.0
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfsAAAHwCAYAAAChTMYRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nsDfGGGOCnIW9McYYE+T+F07FZZ1F40qnAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 576x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_training(history_1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "id": "6En5oYTWcrCH",
    "colab_type": "code",
    "colab": {}
   },
   "outputs": [],
   "source": [
    "model.save(base_dir + 'model/' + 'keras_medicine_tl.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "id": "McFN4FnqIaEn",
    "colab_type": "code",
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 136.0
    },
    "outputId": "3ecfddc5-5869-4fa3-a3f9-f81488898309",
    "executionInfo": {
     "status": "ok",
     "timestamp": 1.555300899858E12,
     "user_tz": -480.0,
     "elapsed": 36833.0,
     "user": {
      "displayName": "J.L H",
      "photoUrl": "https://lh5.googleusercontent.com/-BRX5TWuOHpE/AAAAAAAAAAI/AAAAAAAAAD0/SzepdPSA1LY/s64/photo.jpg",
      "userId": "14158868708361974289"
     }
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[0.68279696 0.2038744  0.11332864]]\n",
      "[INFO] classifying image...\n",
      "[0.68279696 0.2038744  0.11332864]\n",
      "0\n",
      "{'枸杞子': 0, '白术 ': 1, '茯苓': 2}\n",
      "{'枸杞子': 0, '白术': 1, '茯苓': 2}\n",
      "枸杞子\n"
     ]
    }
   ],
   "source": [
    "#预测\n",
    "import numpy as np\n",
    "\n",
    "from keras.models import load_model\n",
    "from keras.preprocessing import image\n",
    "from keras.applications.imagenet_utils import preprocess_input, decode_predictions\n",
    "#from keras.applications.inception_v3 import preprocess_input, decode_predictions\n",
    "\n",
    "#加载模型预测\n",
    "model = load_model(base_dir + 'model/' + 'keras_medicine_tl.h5')\n",
    "\n",
    "# 1.读取图片并处理（方法一）\n",
    "img_path = base_dir + 'tmp/' + 'gq.png'\n",
    "img = image.load_img(img_path, target_size=(224, 224))\n",
    "x = image.img_to_array(img)\n",
    "x /= 255.\n",
    "x = np.expand_dims(x, axis=0)\n",
    "#x = preprocess_input(x) # 使用imagenet_utils的preprocess_input方法处理图片，会导致分类错误,改为 /255\n",
    "preds = model.predict(x)\n",
    "print(preds)\n",
    "\n",
    "\n",
    "print(\"[INFO] classifying image...\")\n",
    "proba = model.predict(x)[0]\n",
    "print(proba)\n",
    "idx = np.argmax(proba)\n",
    "print(idx)\n",
    "print(train_generator.class_indices)\n",
    "print(val_generator.class_indices)\n",
    "\n",
    "inv_map = {v: k for k, v in train_generator.class_indices.items()}\n",
    "label = inv_map[idx]\n",
    "print(label)"
   ]
  }
 ],
 "metadata": {
  "colab": {
   "name": "MedicineTransferLearning.ipynb",
   "version": "0.3.2",
   "provenance": [
    {
     "file_id": "1KcFiw5A-MZI92mDvNnwCjemnAf3ySz0w",
     "timestamp": 1.555253417723E12
    }
   ],
   "collapsed_sections": []
  },
  "kernelspec": {
   "name": "python3",
   "display_name": "Python 3"
  },
  "accelerator": "GPU"
 },
 "nbformat": 4,
 "nbformat_minor": 0
}