DR_detection.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "DR-detection.ipynb",
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "9Llj4gPYfnis"
      },
      "source": [
        "#Imports\n",
        "!pip install kaggle\n",
        "from zipfile import ZipFile as z\n",
        "from google.colab import files"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rDIyNteZfqGX"
      },
      "source": [
        "#Kaggle Data Import\n",
        "files.upload()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "65oBZ8ENfscs"
      },
      "source": [
        "!mkdir -p ~/.kaggle\n",
        "!cp kaggle.json ~/.kaggle/\n",
        "\n",
        "!chmod 600 ~/.kaggle/kaggle.json"
      ],
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZIBPrytYftqo"
      },
      "source": [
        "!kaggle datasets download --force -d tanlikesmath/diabetic-retinopathy-resized"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3bRo3oAOf0uZ"
      },
      "source": [
        "file_name = \"diabetic-retinopathy-resized.zip\"\n",
        "\n",
        "with z(file_name, 'r') as zi:\n",
        "  zi.extractall()\n",
        "  print('done')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FtoJ-m70f2Je"
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "I6LRgi9keldf"
      },
      "source": [
        "import os\n",
        "import pandas as pd \n",
        "import numpy as np\n",
        "import tensorflow as tf\n",
        "import matplotlib.pyplot as plt\n",
        "from glob import glob\n",
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Conv2D, MaxPooling2D, Flatten, Dense, BatchNormalization, Dropout\n",
        "from sklearn.model_selection import train_test_split"
      ],
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8bK6k-40er5u"
      },
      "source": [
        "path = \"/content/resized_train_cropped/resized_train_cropped\"\n",
        "data = \"/content/resized_train_cropped/resized_train_cropped\"\n",
        "print('number of images in total - ',len(os.listdir(data)))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QV6FoI5iiNx9"
      },
      "source": [
        "data = pd.read_csv(\"/content/trainLabels.csv\") \n",
        "data.head()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6O7eL28tmcAz"
      },
      "source": [
        "data['image_name'] = [i+\".jpeg\" for i in data['image'].values]\n",
        "data.head()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XwX2OBkFmhmX"
      },
      "source": [
        "train, val = train_test_split(data, test_size=0.15)"
      ],
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-KAk7G4xmqVA"
      },
      "source": [
        "train.shape, val.shape"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JxUcH9YgmtWp"
      },
      "source": [
        "import cv2\n",
        "def load_ben_color(image):\n",
        "    IMG_SIZE = 224\n",
        "    sigmaX=10\n",
        "    image = cv2.resize(image, (IMG_SIZE, IMG_SIZE))\n",
        "    image=cv2.addWeighted ( image,4, cv2.GaussianBlur( image , (0,0) , sigmaX) ,-4 ,128)\n",
        "    return image"
      ],
      "execution_count": 14,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NyJFu1nSmvJT"
      },
      "source": [
        "data_gen = ImageDataGenerator(rescale=1/255.,\n",
        "                              zoom_range=0.15,\n",
        "                              fill_mode='constant',\n",
        "                              cval=0.,\n",
        "                              horizontal_flip=True,\n",
        "                              vertical_flip=True,\n",
        "                              preprocessing_function=load_ben_color)"
      ],
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VsDrNsQXmw8I"
      },
      "source": [
        "# batch size\n",
        "bs = 128\n",
        "\n",
        "train_gen = data_gen.flow_from_dataframe(train, \n",
        "                                        path,\n",
        "                                         x_col=\"image_name\", y_col=\"level\", class_mode=\"raw\",\n",
        "                                         batch_size=bs,\n",
        "                                         target_size=(224, 224))\n",
        "val_gen = data_gen.flow_from_dataframe(val,\n",
        "                                       path,\n",
        "                                       x_col=\"image_name\", y_col=\"level\", class_mode=\"raw\",\n",
        "                                       batch_size=bs,\n",
        "                                       target_size=(224, 224))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2OhC8pZfmzEX"
      },
      "source": [
        "from keras.applications.resnet50 import ResNet50\n",
        "import keras.layers as L\n",
        "from keras.models import Model"
      ],
      "execution_count": 17,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TZYMkCZKm2AZ"
      },
      "source": [
        "base_model = ResNet50(weights='imagenet',\n",
        "                   include_top=False,\n",
        "                   input_shape=(224, 224, 3))\n",
        "\n",
        "x = base_model.output\n",
        "x = L.GlobalMaxPooling2D()(x)\n",
        "x = L.BatchNormalization()(x)\n",
        "x = L.Dropout(0.2)(x)\n",
        "x = L.Dense(1024, activation=\"relu\")(x)\n",
        "x = L.Dropout(0.1)(x)\n",
        "x = L.Dense(64, activation=\"relu\")(x)\n",
        "predictions = L.Dense(5, activation='softmax')(x)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VsGaxWapnCpY"
      },
      "source": [
        "model = Model(inputs=base_model.input, outputs=predictions)"
      ],
      "execution_count": 19,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GObADzuwnEX9"
      },
      "source": [
        "for layer in base_model.layers[:-20]: layer.trainable = False"
      ],
      "execution_count": 20,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qPX7ocjunGTK"
      },
      "source": [
        "model.compile(optimizer='adam', loss='sparse_categorical_crossentropy', metrics=['accuracy'])"
      ],
      "execution_count": 41,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KuM5as3fnIG9"
      },
      "source": [
        "model.summary()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dOJLjQC7nMld"
      },
      "source": [
        "from keras.callbacks import ModelCheckpoint, ReduceLROnPlateau"
      ],
      "execution_count": 43,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "BpZaRKH3nOqE",
        "outputId": "2f951d90-7897-48c9-9f60-c2d08b389c91"
      },
      "source": [
        "mmodel_history = model.fit(train_gen, steps_per_epoch=int(29842/128), epochs=2,  validation_data=val_gen, validation_steps=int(5266/128))"
      ],
      "execution_count": 45,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/10\n",
            "233/233 [==============================] - 978s 4s/step - loss: 1.1454 - accuracy: 0.6850 - val_loss: 0.9753 - val_accuracy: 0.7348\n",
            "Epoch 2/10\n",
            "233/233 [==============================] - 969s 4s/step - loss: 0.9068 - accuracy: 0.7303 - val_loss: 0.9026 - val_accuracy: 0.7361\n",
            "Epoch 3/10\n",
            "233/233 [==============================] - 962s 4s/step - loss: 0.8799 - accuracy: 0.7325 - val_loss: 0.8784 - val_accuracy: 0.7365\n",
            "Epoch 4/10\n",
            "233/233 [==============================] - 963s 4s/step - loss: 0.8699 - accuracy: 0.7357 - val_loss: 0.8539 - val_accuracy: 0.7361\n",
            "Epoch 5/10\n",
            "233/233 [==============================] - 961s 4s/step - loss: 0.8728 - accuracy: 0.7331 - val_loss: 0.9105 - val_accuracy: 0.7327\n",
            "Epoch 6/10\n",
            "233/233 [==============================] - 947s 4s/step - loss: 0.8668 - accuracy: 0.7306 - val_loss: 0.8627 - val_accuracy: 0.7357\n",
            "Epoch 7/10\n",
            "233/233 [==============================] - 919s 4s/step - loss: 0.8541 - accuracy: 0.7368 - val_loss: 0.8391 - val_accuracy: 0.7359\n",
            "Epoch 8/10\n",
            "233/233 [==============================] - 921s 4s/step - loss: 0.8600 - accuracy: 0.7339 - val_loss: 0.8432 - val_accuracy: 0.7361\n",
            "Epoch 9/10\n",
            "233/233 [==============================] - 916s 4s/step - loss: 0.8454 - accuracy: 0.7397 - val_loss: 0.8417 - val_accuracy: 0.7361\n",
            "Epoch 10/10\n",
            "233/233 [==============================] - 934s 4s/step - loss: 0.8432 - accuracy: 0.7364 - val_loss: 0.8789 - val_accuracy: 0.7355\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 577
        },
        "id": "3RodgrSlsJ4Z",
        "outputId": "8ea870bb-9967-41d7-996e-72a956fa533b"
      },
      "source": [
        "#SAVING WEIGHTS\n",
        "\n",
        "model.save('dr_weights.h5')\n",
        "\n",
        "#MODEL\n",
        "\n",
        "accNT = mmodel_history.history['accuracy']\n",
        "val_accNT = mmodel_history.history['val_accuracy']\n",
        "lossNT = mmodel_history.history['loss']\n",
        "val_lossNT = mmodel_history.history['val_loss']\n",
        "epochsNT = range(len(accNT))\n",
        "\n",
        "plt.plot(epochsNT, accNT, 'r', label='Training accuracy')\n",
        "plt.plot(epochsNT, val_accNT, 'b', label='Validation accuracy')\n",
        "plt.title('Training and validation accuracy')\n",
        "plt.legend(loc=0)\n",
        "plt.figure()\n",
        "\n",
        "plt.show()\n",
        "\n",
        "plt.plot(epochsNT, lossNT, 'r', label='Training loss')\n",
        "plt.plot(epochsNT, val_lossNT, 'b', label='Validation loss')\n",
        "plt.title('Training and validation loss')\n",
        "plt.legend(loc=0)\n",
        "plt.figure()\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": 46,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 0 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 0 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    }
  ]
}