DIP_041.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "DIP_041.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/Dr-Carlos-Villasenor/DIP/blob/main/DIP_041.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4zp_0_xZvw4K"
      },
      "source": [
        "# Análisis y procesamiento de imágenes\n",
        "## Dr. Carlos Villaseñor\n",
        "### DIP_041 - Transferencia de aprendizaje"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qMJem6ZuwknJ"
      },
      "source": [
        "Importar bibliotecas\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mtXnehErvsnR"
      },
      "source": [
        "import pandas as pd\n",
        "import numpy as np\n",
        "import os\n",
        "import keras\n",
        "import matplotlib.pyplot as plt\n",
        "from keras.layers import Dense,GlobalAveragePooling2D\n",
        "from keras.applications import mobilenet\n",
        "from keras.preprocessing import image\n",
        "from keras.applications.mobilenet import preprocess_input\n",
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "from keras.callbacks import ModelCheckpoint\n",
        "from keras.models import Model"
      ],
      "execution_count": 80,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Para subir los archivos tenemos dos opciones, cargar su carpeta de drive:"
      ],
      "metadata": {
        "id": "Wh-J-B3B0g3h"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "metadata": {
        "id": "xdnOjKX_0gWw"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4-39u5Pp1Lkj"
      },
      "source": [
        "Cargar un archivo .zip y descomprimirlo\n",
        "\n",
        "- Carge el archivo 'TL_dataset.zip' en la pestaña archivos en el lateral izquierdo\n",
        "- Corra la siguiente linea para descomprimir los archivos"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fwbljBNp1hKK"
      },
      "source": [
        "!unzip TL_dataset.zip"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "shvPrpcuw43_"
      },
      "source": [
        "Importar MobileNet base pero descartaremos las últimas capa"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4Q6iGl62xPNJ"
      },
      "source": [
        "base_model=mobilenet.MobileNet(weights='imagenet',include_top=False)\n",
        "print(base_model.summary())"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "aCeLxtlEyCJV"
      },
      "source": [
        "Crear nueva Red Neuronal"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AYQX0UAbyHzA"
      },
      "source": [
        "x=base_model.output\n",
        "x=GlobalAveragePooling2D()(x)\n",
        "x=Dense(64,activation='relu')(x) \n",
        "x=Dense(32,activation='relu')(x) \n",
        "preds=Dense(3, activation='softmax')(x)"
      ],
      "execution_count": 82,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "b9fBbvIzzJG6"
      },
      "source": [
        "Pegar Redes Neuronales"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PkdA8UWWzSrS"
      },
      "source": [
        "model=Model(inputs=base_model.input, outputs=preds)\n",
        "print(model.summary())"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5Sg7ZRB0zq-i"
      },
      "source": [
        "Definir que capas se entrenan"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OGYwhdoFzqVk"
      },
      "source": [
        "for layer in model.layers[:-5]:\n",
        "    layer.trainable=False\n",
        "for layer in model.layers[-5:]:\n",
        "    layer.trainable=True"
      ],
      "execution_count": 84,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qKTX6gWt14cy"
      },
      "source": [
        "Crear generador desde carpeta"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "p7oy8_sZ2Mc-"
      },
      "source": [
        "train_datagen=ImageDataGenerator(zoom_range=0.2, horizontal_flip=True,\n",
        "                                 width_shift_range=0.2,\n",
        "                                 height_shift_range=0.2,\n",
        "                                 preprocessing_function=preprocess_input)\n",
        "\n",
        "train_generator=train_datagen.flow_from_directory('./TL_dataset/train/',\n",
        "                                                 target_size=(224,224),\n",
        "                                                 color_mode='rgb',\n",
        "                                                 batch_size=32,\n",
        "                                                 class_mode='categorical',\n",
        "                                                 shuffle=True)\n",
        "\n",
        "test_datagen=ImageDataGenerator(preprocessing_function=preprocess_input)\n",
        "\n",
        "test_generator=test_datagen.flow_from_directory('./TL_dataset/test/',\n",
        "                                                 target_size=(224,224),\n",
        "                                                 color_mode='rgb',\n",
        "                                                 batch_size=31,\n",
        "                                                 class_mode='categorical',\n",
        "                                                 shuffle=False)\n",
        "\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "F9zw5rV33psh"
      },
      "source": [
        "Compilar Modelo y entrenar desde carpeta"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "efD515JO3zlL"
      },
      "source": [
        "model.compile(optimizer='Adam',loss='categorical_crossentropy',\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "step_size_train=train_generator.n//train_generator.batch_size\n",
        "\n",
        "model.fit_generator(generator=train_generator,\n",
        "                    steps_per_epoch=step_size_train,\n",
        "                    epochs=6)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Evaluación final del modelo y gruadado del mismo"
      ],
      "metadata": {
        "id": "2MqTO1_b4QkY"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "step_size_test = test_generator.n // test_generator.batch_size\n",
        "model.evaluate(test_generator, steps=step_size_test)\n",
        "\n",
        "model.save('model.hdf5')"
      ],
      "metadata": {
        "id": "GnjOuMwev8Tg"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Solo nos queda probar el modelo"
      ],
      "metadata": {
        "id": "pm4Fic9A57Yi"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from keras.models import load_model\n",
        "model = load_model('./model.hdf5')\n",
        "\n",
        "test_datagen = ImageDataGenerator(preprocessing_function=preprocess_input)\n",
        "test_generator = test_datagen.flow_from_directory('./TL_dataset/test/',\n",
        "                                                  target_size=(224, 224),\n",
        "                                                  color_mode='rgb',\n",
        "                                                  batch_size=1,\n",
        "                                                  class_mode='categorical',\n",
        "                                                  shuffle=True)"
      ],
      "metadata": {
        "id": "Mbx5dggL6449"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt \n",
        "image, classe = next(test_generator)\n",
        "classes = np.argmax(classe, axis=-1)\n",
        "pred = np.argmax(model.predict(image), axis=1)\n",
        "\n",
        "names = {0:'cat', 1:'dog', 2:'horse'}\n",
        "\n",
        "plt.imshow(image[0])\n",
        "plt.title('Network pred: ' + names[pred[0]])"
      ],
      "metadata": {
        "id": "yEDoMFLt7VLi"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}