densenet.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "import os\n",
        "import matplotlib.pyplot as plt\n",
        "import tensorflow as tf\n",
        "\n",
        "from keras.utils.np_utils import to_categorical\n",
        "from keras.preprocessing.image import load_img, img_to_array\n",
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "\n",
        "from sklearn.metrics import classification_report, log_loss, accuracy_score\n",
        "from sklearn.model_selection import train_test_split"
      ],
      "metadata": {
        "id": "lSQLppgTyYOD"
      },
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "directory = '/content/Fish_Dataset/Fish_Dataset'"
      ],
      "metadata": {
        "id": "Cx_9WEjeyZ8b"
      },
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "Name=[]\n",
        "for file in os.listdir(directory):\n",
        "    if file[-4:]!='pt.m' and file[-4:]!='.txt':\n",
        "        Name+=[file]\n",
        "print(Name)\n",
        "print(len(Name))\n",
        "['Hourse Mackerel', 'Black Sea Sprat', 'Sea Bass', 'Red Mullet', 'Trout', 'Striped Red Mullet', 'Shrimp', 'Gilt-Head Bream', 'Red Sea Bream']\n",
        "9\n",
        "N=[]\n",
        "for i in range(len(Name)):\n",
        "    N+=[i]\n",
        "    \n",
        "normal_mapping=dict(zip(Name,N)) \n",
        "reverse_mapping=dict(zip(N,Name)) \n",
        "\n",
        "def mapper(value):\n",
        "    return reverse_mapping[value]\n",
        "File=[]\n",
        "for file in os.listdir(directory):\n",
        "    if file[-4:]!='pt.m' and file[-4:]!='.txt':\n",
        "        File+=[file]\n",
        "        print(file)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZkLG1N9RyeLy",
        "outputId": "8ed04f1d-6baa-4183-b90c-ce0e2472f064"
      },
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "['Red Mullet', 'Sea Bass', 'Striped Red Mullet', 'Hourse Mackerel', 'Trout', 'Black Sea Sprat', 'Shrimp', 'Gilt-Head Bream', 'Red Sea Bream']\n",
            "9\n",
            "Red Mullet\n",
            "Sea Bass\n",
            "Striped Red Mullet\n",
            "Hourse Mackerel\n",
            "Trout\n",
            "Black Sea Sprat\n",
            "Shrimp\n",
            "Gilt-Head Bream\n",
            "Red Sea Bream\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "dataset=[]\n",
        "testset=[]\n",
        "count=0\n",
        "for file in File:\n",
        "    path=os.path.join(directory,file,file)\n",
        "    t=0\n",
        "    for im in os.listdir(path):\n",
        "        if im[-4:]!='pt.m' and im[-4:]!='.txt':\n",
        "            image=load_img(os.path.join(path,im), grayscale=False, color_mode='rgb', target_size=(100,100))\n",
        "            image=img_to_array(image)\n",
        "            image=image/255.0\n",
        "            if t<=800:\n",
        "                dataset.append([image,count])\n",
        "            else:   \n",
        "                testset.append([image,count])\n",
        "            t+=1\n",
        "    count=count+1"
      ],
      "metadata": {
        "id": "Y4qZTCyqzJy6"
      },
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "data,labels0=zip(*dataset)\n",
        "test,tlabels0=zip(*testset)"
      ],
      "metadata": {
        "id": "X722NquczNEv"
      },
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "labels1=to_categorical(labels0)\n",
        "data=np.array(data)\n",
        "labels=np.array(labels1)"
      ],
      "metadata": {
        "id": "4d4_KaXRzOmo"
      },
      "execution_count": 17,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "tlabels1=to_categorical(tlabels0)\n",
        "test=np.array(test)\n",
        "tlabels=np.array(tlabels1)"
      ],
      "metadata": {
        "id": "RxNlMyA3zQKf"
      },
      "execution_count": 18,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "trainx,testx,trainy,testy=train_test_split(data,labels,test_size=0.2,random_state=44)"
      ],
      "metadata": {
        "id": "QC2waqH7zRtW"
      },
      "execution_count": 19,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(trainx.shape)\n",
        "print(testx.shape)\n",
        "print(trainy.shape)\n",
        "print(testy.shape)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "wymaM3hHzTDn",
        "outputId": "26b6683f-4c39-487c-a3c5-ce7cce1907a9"
      },
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(5767, 100, 100, 3)\n",
            "(1442, 100, 100, 3)\n",
            "(5767, 9)\n",
            "(1442, 9)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "datagen = ImageDataGenerator(horizontal_flip=True,vertical_flip=True,rotation_range=20,zoom_range=0.2,\n",
        "                        width_shift_range=0.2,height_shift_range=0.2,shear_range=0.1,fill_mode=\"nearest\")\n",
        "pretrained_model3 = tf.keras.applications.DenseNet201(input_shape=(100,100,3),include_top=False,weights='imagenet',pooling='avg')\n",
        "pretrained_model3.trainable = False\n",
        "inputs3 = pretrained_model3.input\n",
        "x3 = tf.keras.layers.Dense(128, activation='relu')(pretrained_model3.output)\n",
        "outputs3 = tf.keras.layers.Dense(9, activation='softmax')(x3)\n",
        "model = tf.keras.Model(inputs=inputs3, outputs=outputs3)\n",
        "model.compile(optimizer='adam',loss='categorical_crossentropy',metrics=['accuracy'])\n",
        "his=model.fit(datagen.flow(trainx,trainy,batch_size=32),validation_data=(testx,testy),epochs=20)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3tIWruzXzVns",
        "outputId": "608a68c9-2754-44d5-c3ba-f50825f18c38"
      },
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/densenet/densenet201_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
            "74842112/74836368 [==============================] - 2s 0us/step\n",
            "74850304/74836368 [==============================] - 2s 0us/step\n",
            "Epoch 1/20\n",
            "181/181 [==============================] - 46s 134ms/step - loss: 0.6516 - accuracy: 0.7734 - val_loss: 0.2619 - val_accuracy: 0.9071\n",
            "Epoch 2/20\n",
            "181/181 [==============================] - 20s 111ms/step - loss: 0.2738 - accuracy: 0.9038 - val_loss: 0.2465 - val_accuracy: 0.9064\n",
            "Epoch 3/20\n",
            "181/181 [==============================] - 19s 105ms/step - loss: 0.2160 - accuracy: 0.9208 - val_loss: 0.1060 - val_accuracy: 0.9653\n",
            "Epoch 4/20\n",
            "181/181 [==============================] - 19s 103ms/step - loss: 0.1720 - accuracy: 0.9398 - val_loss: 0.1459 - val_accuracy: 0.9445\n",
            "Epoch 5/20\n",
            "181/181 [==============================] - 18s 102ms/step - loss: 0.1670 - accuracy: 0.9404 - val_loss: 0.0864 - val_accuracy: 0.9660\n",
            "Epoch 6/20\n",
            "181/181 [==============================] - 19s 106ms/step - loss: 0.1463 - accuracy: 0.9456 - val_loss: 0.2525 - val_accuracy: 0.9223\n",
            "Epoch 7/20\n",
            "181/181 [==============================] - 21s 114ms/step - loss: 0.1435 - accuracy: 0.9483 - val_loss: 0.0790 - val_accuracy: 0.9764\n",
            "Epoch 8/20\n",
            "181/181 [==============================] - 19s 102ms/step - loss: 0.1325 - accuracy: 0.9516 - val_loss: 0.1322 - val_accuracy: 0.9521\n",
            "Epoch 9/20\n",
            "181/181 [==============================] - 18s 102ms/step - loss: 0.1280 - accuracy: 0.9547 - val_loss: 0.0728 - val_accuracy: 0.9750\n",
            "Epoch 10/20\n",
            "181/181 [==============================] - 19s 106ms/step - loss: 0.1159 - accuracy: 0.9580 - val_loss: 0.0822 - val_accuracy: 0.9743\n",
            "Epoch 11/20\n",
            "181/181 [==============================] - 18s 102ms/step - loss: 0.1172 - accuracy: 0.9587 - val_loss: 0.1070 - val_accuracy: 0.9598\n",
            "Epoch 12/20\n",
            "181/181 [==============================] - 19s 105ms/step - loss: 0.1271 - accuracy: 0.9572 - val_loss: 0.1643 - val_accuracy: 0.9445\n",
            "Epoch 13/20\n",
            "181/181 [==============================] - 19s 102ms/step - loss: 0.1008 - accuracy: 0.9650 - val_loss: 0.0760 - val_accuracy: 0.9730\n",
            "Epoch 14/20\n",
            "181/181 [==============================] - 19s 103ms/step - loss: 0.1189 - accuracy: 0.9565 - val_loss: 0.1356 - val_accuracy: 0.9563\n",
            "Epoch 15/20\n",
            "181/181 [==============================] - 19s 106ms/step - loss: 0.1050 - accuracy: 0.9636 - val_loss: 0.0749 - val_accuracy: 0.9688\n",
            "Epoch 16/20\n",
            "181/181 [==============================] - 19s 102ms/step - loss: 0.0983 - accuracy: 0.9651 - val_loss: 0.0859 - val_accuracy: 0.9695\n",
            "Epoch 17/20\n",
            "181/181 [==============================] - 19s 106ms/step - loss: 0.1018 - accuracy: 0.9641 - val_loss: 0.1369 - val_accuracy: 0.9480\n",
            "Epoch 18/20\n",
            "181/181 [==============================] - 19s 105ms/step - loss: 0.0848 - accuracy: 0.9705 - val_loss: 0.0575 - val_accuracy: 0.9792\n",
            "Epoch 19/20\n",
            "181/181 [==============================] - 19s 105ms/step - loss: 0.1009 - accuracy: 0.9655 - val_loss: 0.0816 - val_accuracy: 0.9771\n",
            "Epoch 20/20\n",
            "181/181 [==============================] - 19s 105ms/step - loss: 0.1013 - accuracy: 0.9624 - val_loss: 0.0535 - val_accuracy: 0.9847\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "y_pred=model.predict(testx)\n",
        "pred=np.argmax(y_pred,axis=1)\n",
        "ground = np.argmax(testy,axis=1)\n",
        "print(classification_report(ground,pred))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vNPKHayFzid7",
        "outputId": "c9a9f66c-9512-4465-c16d-c451ccaa2e1b"
      },
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       1.00      0.99      0.99       134\n",
            "           1       0.99      0.98      0.98       161\n",
            "           2       0.98      0.96      0.97       169\n",
            "           3       0.99      1.00      1.00       170\n",
            "           4       0.99      0.99      0.99       163\n",
            "           5       0.99      0.98      0.99       172\n",
            "           6       0.96      1.00      0.98       138\n",
            "           7       0.98      0.98      0.98       168\n",
            "           8       0.98      0.99      0.99       167\n",
            "\n",
            "    accuracy                           0.98      1442\n",
            "   macro avg       0.98      0.99      0.98      1442\n",
            "weighted avg       0.98      0.98      0.98      1442\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "get_acc = his.history['accuracy']\n",
        "value_acc = his.history['val_accuracy']\n",
        "get_loss = his.history['loss']\n",
        "validation_loss = his.history['val_loss']\n",
        "\n",
        "epochs = range(len(get_acc))\n",
        "plt.plot(epochs, get_acc, 'r', label='Accuracy of Training data')\n",
        "plt.plot(epochs, value_acc, 'b', label='Accuracy of Validation data')\n",
        "plt.title('Training vs validation accuracy')\n",
        "plt.legend(loc=0)\n",
        "plt.figure()\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 298
        },
        "id": "w8BYgNMFzk7A",
        "outputId": "6a59ecc6-c67f-4ab2-90ed-31f84ffa3aea"
      },
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 0 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "epochs = range(len(get_loss))\n",
        "plt.plot(epochs, get_loss, 'r', label='Loss of Training data')\n",
        "plt.plot(epochs, validation_loss, 'b', label='Loss of Validation data')\n",
        "plt.title('Training vs validation loss')\n",
        "plt.legend(loc=0)\n",
        "plt.figure()\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 298
        },
        "id": "aqgy8CD1zmbv",
        "outputId": "5ffe1b6b-a2da-4fde-9d9b-a902cc4a3c40"
      },
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 0 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "image = load_img(\"/content/Fish_Dataset/Fish_Dataset/Red Mullet/Red Mullet/00003.png\",target_size=(100,100))"
      ],
      "metadata": {
        "id": "HfsX4AMpzt0u"
      },
      "execution_count": 26,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "image=img_to_array(image) \n",
        "image=image/255.0\n",
        "prediction_image=np.array(image)\n",
        "prediction_image= np.expand_dims(image, axis=0)"
      ],
      "metadata": {
        "id": "2rGrutp0zy2l"
      },
      "execution_count": 27,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "prediction=model.predict(prediction_image)\n",
        "value=np.argmax(prediction)\n",
        "move_name=mapper(value)\n",
        "#print(prediction)\n",
        "#print(value)\n",
        "print(\"Prediction is {}.\".format(move_name))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "p_g3F0fuz0qj",
        "outputId": "747e3fd4-d5f7-4f4f-9d79-d9989dc9ac45"
      },
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Prediction is Red Mullet.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(test.shape)\n",
        "prediction2=model.predict(test)\n",
        "print(prediction2.shape)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "j99bOdqkz2ej",
        "outputId": "82365610-ddd8-4578-89a9-4fc2dd9644ef"
      },
      "execution_count": 29,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(1791, 100, 100, 3)\n",
            "(1791, 9)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "PRED=[]\n",
        "for item in prediction2:\n",
        "    value2=np.argmax(item)      \n",
        "    PRED+=[value2]"
      ],
      "metadata": {
        "id": "UpTVhnf4z3zp"
      },
      "execution_count": 30,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "ANS=tlabels0\n",
        "accuracy=accuracy_score(ANS,PRED)\n",
        "print(accuracy)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dKrNr-GMz5-p",
        "outputId": "1034ef5e-7e0e-4ef2-ac10-f8e8b75efc07"
      },
      "execution_count": 31,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0.9821328866554997\n"
          ]
        }
      ]
    }
  ]
}