AV-Intel-Scene-Classification-Challenge/scene-classification-final-code.ipynb

{"nbformat":4,"nbformat_minor":0,"metadata":{"colab":{"name":"scene-classification-final-code.ipynb","version":"0.3.2","provenance":[],"collapsed_sections":[]},"kernelspec":{"name":"python3","display_name":"Python 3"},"accelerator":"GPU"},"cells":[{"metadata":{"id":"ZgiJmF-3pe3s","colab_type":"text"},"cell_type":"markdown","source":["{'buildings' -> 0, \n","\n","'forest' -> 1,\n","\n","'glacier' -> 2,\n","\n","'mountain' -> 3,\n","\n","'sea' -> 4,\n","\n","'street' -> 5 }"]},{"metadata":{"id":"n3RYV4j1pM_s","colab_type":"text"},"cell_type":"markdown","source":[""]},{"metadata":{"id":"Z0Clg3m3nxeM","colab_type":"code","colab":{}},"cell_type":"code","source":["labels = ['buildings', 'forest', 'glacier', 'mountain', 'sea', 'street']"],"execution_count":0,"outputs":[]},{"metadata":{"id":"XPLyRQHKmdUq","colab_type":"code","colab":{}},"cell_type":"code","source":["#from google.colab import drive\n","#drive.mount('/content/gdrive')"],"execution_count":0,"outputs":[]},{"metadata":{"_uuid":"f2b1b45d7222c7109884df0029385d81718f7209","id":"CNui6my2ipD2","colab_type":"code","colab":{}},"cell_type":"code","source":["import pandas as pd\n","import numpy as np\n","import matplotlib.pyplot as plt\n","import json\n","import os\n","from tqdm import tqdm, tqdm_notebook\n","\n","import tensorflow as tf\n","from tensorflow.keras.models import Sequential, Model\n","from tensorflow.keras.layers import *\n","from tensorflow.keras.optimizers import *\n","from tensorflow.keras.applications import *\n","from tensorflow.keras.callbacks import *\n","from tensorflow.keras.initializers import *\n","from tensorflow.keras.preprocessing.image import ImageDataGenerator"],"execution_count":0,"outputs":[]},{"metadata":{"_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","id":"mK84XzEvipEJ","colab_type":"code","colab":{}},"cell_type":"code","source":["# Set home directory of this project\n","home_dir = '/content/gdrive/My Drive/data-science-my-projects/AV-Intel-Scene-Classification-Challenge'"],"execution_count":0,"outputs":[]},{"metadata":{"_uuid":"4d3374968aa7f3881f1b83408f272367f0e63619","id":"cwFAsswOipEY","colab_type":"text"},"cell_type":"markdown","source":["## Read and set up data"]},{"metadata":{"_cell_guid":"79c7e3d0-c299-4dcb-8224-4455121ee9b0","_uuid":"d629ff2d2480ee46fbb7e2d37f6b5fab8052498a","id":"e-JCRYi3ipEb","colab_type":"code","colab":{}},"cell_type":"code","source":["# Read data\n","dataset_dir = os.path.join(home_dir, \"dataset\")\n","train_dir = os.path.join(dataset_dir, \"train\")\n","\n","\n","train_df = pd.read_csv(dataset_dir + '/train.csv')\n","train_df.head()"],"execution_count":0,"outputs":[]},{"metadata":{"_uuid":"2ffaa5aba1e6197e0f2ced15187c06aa87cb4726","id":"BXPM-y_kipEo","colab_type":"code","colab":{}},"cell_type":"code","source":["# Read and display an image\n","image = plt.imread(os.path.join(train_dir, os.listdir(train_dir)[25]))\n","print(\"Image shape =\", image.shape)\n","train_input_shape = image.shape\n","plt.imshow(image)\n","plt.show()"],"execution_count":0,"outputs":[]},{"metadata":{"id":"B7VO2ODbtTeG","colab_type":"code","colab":{}},"cell_type":"code","source":["# Number of unique classes\n","n_classes = len(train_df.label.unique())\n","print(\"Number of unique classes =\", n_classes)"],"execution_count":0,"outputs":[]},{"metadata":{"_uuid":"9afc1b8533ad9c6167e3a50432127e4c6e9f44e0","id":"R7b2hbCpipEy","colab_type":"text"},"cell_type":"markdown","source":["## Image Augmentation"]},{"metadata":{"_uuid":"5bdbbb3b523d3e82c9c83508842c8536ddb645a9","id":"Hf2H6y0ZipE1","colab_type":"code","outputId":"6e1aa3da-d104-4cdf-e751-dd51ce3177f9","executionInfo":{"status":"ok","timestamp":1554030294746,"user_tz":-330,"elapsed":82133,"user":{"displayName":"Sup H","photoUrl":"","userId":"04337592286741421479"}},"colab":{"base_uri":"https://localhost:8080/","height":68}},"cell_type":"code","source":["# Augment data\n","batch_size = 32\n","#train_input_shape = (64, 64, 3)\n","\n","train_datagen = ImageDataGenerator(validation_split=0.2,\n","                                   rescale=1./255.,\n","                                   rotation_range=45,\n","                                   width_shift_range=0.2,\n","                                   height_shift_range=0.2,\n","                                   shear_range=0.7,\n","                                   zoom_range=0.7,\n","                                   horizontal_flip=True,\n","                                   #vertical_flip=True,\n","                                  )\n","\n","train_generator = train_datagen.flow_from_dataframe(dataframe=train_df, \n","                                                    directory=train_dir, \n","                                                    x_col=\"image_name\", \n","                                                    y_col=\"label\", \n","                                                    class_mode=\"other\",\n","                                                    subset=\"training\",\n","                                                    target_size=train_input_shape[0:2],\n","                                                    shuffle=True,\n","                                                    batch_size=batch_size)\n","\n","valid_generator = train_datagen.flow_from_dataframe(dataframe=train_df, \n","                                                    directory=train_dir, \n","                                                    x_col=\"image_name\", \n","                                                    y_col=\"label\", \n","                                                    class_mode=\"other\",\n","                                                    subset=\"validation\",\n","                                                    target_size=train_input_shape[0:2],\n","                                                    shuffle=True,\n","                                                    batch_size=batch_size)\n","\n","STEP_SIZE_TRAIN = train_generator.n//train_generator.batch_size\n","STEP_SIZE_VALID = valid_generator.n//valid_generator.batch_size\n","print(\"Total number of batches =\", STEP_SIZE_TRAIN, \"and\", STEP_SIZE_VALID)"],"execution_count":0,"outputs":[{"output_type":"stream","text":["Found 13628 images.\n","Found 3406 images.\n","Total number of batches = 425 and 106\n"],"name":"stdout"}]},{"metadata":{"_uuid":"14162cea761f425a19a7f1a0589e445c8f0b06c7","id":"FUjYGRNdipFB","colab_type":"text"},"cell_type":"markdown","source":["## Build model"]},{"metadata":{"_uuid":"0015c78383b92457b131f2f79379847503f01833","id":"vBmJPc8yipFE","colab_type":"code","colab":{}},"cell_type":"code","source":["# Load pre-trained model\n","resnet50 = ResNet50(weights='imagenet', include_top=False, input_shape=train_input_shape)\n","\n","# Add layers at the end\n","X = resnet50.output\n","X = Flatten()(X)\n","\n","X = Dense(16, kernel_initializer='he_uniform')(X)\n","X = Dropout(0.5)(X)\n","X = BatchNormalization()(X)\n","X = Activation('relu')(X)\n","\n","output = Dense(n_classes, activation='softmax')(X)\n","\n","model = Model(inputs=resnet50.input, outputs=output)\n","\n","#model.summary()"],"execution_count":0,"outputs":[]},{"metadata":{"_uuid":"6ce184ce000b658fc951097822dc1c5605152d47","id":"xPxVLN3YipFk","colab_type":"code","colab":{}},"cell_type":"code","source":["# Train all layers \n","optimizer=Adam(lr=0.0001)\n","\n","model.compile(loss='sparse_categorical_crossentropy',\n","              optimizer=optimizer, \n","              metrics=['accuracy'])\n","\n","\n","early_stop = EarlyStopping(monitor='val_acc', patience=20, verbose=1, \n","                           mode='auto', restore_best_weights=True)\n","\n","reduce_lr = ReduceLROnPlateau(monitor='val_loss', factor=0.1, patience=5, \n","                              verbose=1, mode='auto')\n","\n","n_epoch = 50\n","\n","history = model.fit_generator(generator=train_generator, steps_per_epoch=STEP_SIZE_TRAIN,\n","                              validation_data=valid_generator, validation_steps=STEP_SIZE_VALID,\n","                              epochs=n_epoch,\n","                              shuffle=True,\n","                              verbose=1,\n","                              callbacks=[early_stop, reduce_lr],\n","                              use_multiprocessing=True,\n","                              workers=10\n","                             )"],"execution_count":0,"outputs":[]},{"metadata":{"id":"7shZsZvp8_l8","colab_type":"code","colab":{}},"cell_type":"code","source":["# Train only last few layers\n","optimizer=Adam(lr=0.0001)\n","\n","for layer in model.layers[:-6]:\n","    layer.trainable = False\n","    \n","model.compile(loss='sparse_categorical_crossentropy',\n","              optimizer=optimizer, \n","              metrics=['accuracy'])\n","\n","\n","early_stop = EarlyStopping(monitor='val_acc', patience=20, verbose=1, \n","                           mode='auto', restore_best_weights=True)\n","\n","reduce_lr = ReduceLROnPlateau(monitor='val_loss', factor=0.1, patience=5, \n","                              verbose=1, mode='auto')\n","\n","n_epoch = 50\n","\n","history = model.fit_generator(generator=train_generator, steps_per_epoch=STEP_SIZE_TRAIN,\n","                              validation_data=valid_generator, validation_steps=STEP_SIZE_VALID,\n","                              epochs=n_epoch,\n","                              shuffle=True,\n","                              verbose=1,\n","                              callbacks=[early_stop, reduce_lr],\n","                              use_multiprocessing=True,\n","                              workers=10\n","                             )"],"execution_count":0,"outputs":[]},{"metadata":{"id":"x8uIuuit8paW","colab_type":"code","colab":{}},"cell_type":"code","source":["# Save the model and weights\n","model_dir = os.path.join(home_dir, \"models\")\n","model.save(model_dir + '/resnet50_model.h5')\n","model.save_weights(model_dir + '/resnet50_weights.h5')"],"execution_count":0,"outputs":[]},{"metadata":{"_uuid":"898075f768db94965133df52fe650334f744541d","id":"8-bDEvFOipF-","colab_type":"code","colab":{}},"cell_type":"code","source":["# Plot the training graph\n","def plot_training(history):\n","    acc = history.history['acc']\n","    val_acc = history.history['val_acc']\n","    loss = history.history['loss']\n","    val_loss = history.history['val_loss']\n","    epochs = range(len(acc))\n","\n","    fig, axes = plt.subplots(1, 2, figsize=(15,5))\n","    \n","    axes[0].plot(epochs, acc, 'r-', label='Training Accuracy')\n","    axes[0].plot(epochs, val_acc, 'b--', label='Validation Accuracy')\n","    axes[0].set_title('Training and Validation Accuracy')\n","    axes[0].legend(loc='best')\n","\n","    axes[1].plot(epochs, loss, 'r-', label='Training Loss')\n","    axes[1].plot(epochs, val_loss, 'b--', label='Validation Loss')\n","    axes[1].set_title('Training and Validation Loss')\n","    axes[1].legend(loc='best')\n","    \n","    plt.show()\n","    \n","plot_training(history)"],"execution_count":0,"outputs":[]},{"metadata":{"_uuid":"cd4bf4945a12d11ce13c9cda6befe55c392c0d8d","id":"NQtv0MCgipGI","colab_type":"code","outputId":"04329264-9589-4f70-ca7e-39500b40533c","executionInfo":{"status":"ok","timestamp":1554054245909,"user_tz":-330,"elapsed":1020844,"user":{"displayName":"Sup H","photoUrl":"","userId":"04337592286741421479"}},"colab":{"base_uri":"https://localhost:8080/","height":51}},"cell_type":"code","source":["# Evaluate on validation set\n","result = model.evaluate_generator(generator=valid_generator, verbose=1)\n","result"],"execution_count":0,"outputs":[{"output_type":"stream","text":["107/107 [==============================] - 40s 376ms/step - loss: 0.3636 - acc: 0.9151\n"],"name":"stdout"},{"output_type":"execute_result","data":{"text/plain":["[0.36362757830158154, 0.91514975]"]},"metadata":{"tags":[]},"execution_count":38}]},{"metadata":{"_uuid":"cd427b2643d0a1e94227977ecf9b579415f9f7d4","id":"ATjwrbViipGR","colab_type":"code","outputId":"e487afb6-7e62-49fe-fe9b-fa2a6af37c99","executionInfo":{"status":"ok","timestamp":1553956573625,"user_tz":-330,"elapsed":5079930,"user":{"displayName":"Sup H","photoUrl":"","userId":"04337592286741421479"}},"colab":{"base_uri":"https://localhost:8080/","height":597}},"cell_type":"code","source":["# Classification report and confusion matrix\n","from sklearn.metrics import *\n","import seaborn as sns\n","\n","def showClassficationReport_Generator(model, valid_generator, STEP_SIZE_VALID):\n","    # Loop on each generator batch and predict\n","    y_pred, y_true = [], []\n","    for i in range(STEP_SIZE_VALID):\n","        (X,y) = next(valid_generator)\n","        y_pred.append(model.predict(X))\n","        y_true.append(y)\n","    \n","    # Create a flat list for y_true and y_pred\n","    y_pred = [subresult for result in y_pred for subresult in result]\n","    y_true = [subresult for result in y_true for subresult in result]\n","    \n","    y_true = np.asarray(y_true).ravel()\n","    \n","    # Update Prediction vector based on argmax\n","    #y_pred = np.asarray(y_pred).astype('float32').ravel()\n","    #y_pred = y_pred >= 0.5\n","    #y_pred = y_pred.astype('int').ravel()\n","    y_pred = np.argmax(y_pred, axis=1)\n","    y_pred = np.asarray(y_pred).ravel()\n","    \n","    # Confusion Matrix\n","    conf_matrix = confusion_matrix(y_true, y_pred, labels=[0,1,2,3,4,5])\n","    sns.heatmap(conf_matrix, annot=True, fmt=\"d\", square=True, cbar=False, \n","                cmap=plt.cm.gray, xticklabels=labels, yticklabels=labels)\n","    plt.ylabel('Actual')\n","    plt.xlabel('Predicted')\n","    plt.title('Confusion Matrix')\n","    plt.show()\n","    \n","    print(classification_report(y_true, y_pred))\n","    #print(\"\\nAUC: \", roc_auc_score(y_true, y_pred, average='micro'))\n","\n","showClassficationReport_Generator(model, valid_generator, STEP_SIZE_VALID)"],"execution_count":0,"outputs":[{"output_type":"display_data","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAVgAAAFnCAYAAADjQFE4AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nKwiCYCMiwQqCINiISLCCIAg28h+pqgCCmNw4XwAAAABJRU5ErkJggg==\n","text/plain":["<Figure size 576x396 with 1 Axes>"]},"metadata":{"tags":[]}},{"output_type":"stream","text":["              precision    recall  f1-score   support\n","\n","           0       0.91      0.92      0.91       507\n","           1       0.99      0.97      0.98       541\n","           2       0.90      0.84      0.87       594\n","           3       0.85      0.91      0.87       600\n","           4       0.94      0.95      0.94       561\n","           5       0.93      0.93      0.93       587\n","\n","   micro avg       0.92      0.92      0.92      3390\n","   macro avg       0.92      0.92      0.92      3390\n","weighted avg       0.92      0.92      0.92      3390\n","\n"],"name":"stdout"}]},{"metadata":{"_uuid":"c841e3ab0b59e0a3bf9ae1885d99533b05f06fc4","id":"YQSdReezipGd","colab_type":"text"},"cell_type":"markdown","source":["## Prepare data for prediction on test set"]},{"metadata":{"id":"6e7bQQBIqtBu","colab_type":"code","outputId":"3a0434e6-8191-479a-eddc-afce7db257d9","executionInfo":{"status":"ok","timestamp":1554054245913,"user_tz":-330,"elapsed":1016410,"user":{"displayName":"Sup H","photoUrl":"","userId":"04337592286741421479"}},"colab":{"base_uri":"https://localhost:8080/","height":34}},"cell_type":"code","source":["test_df = pd.read_csv(dataset_dir + \"/test_WyRytb0.csv\")\n","test_df.shape"],"execution_count":0,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(7301, 1)"]},"metadata":{"tags":[]},"execution_count":39}]},{"metadata":{"_uuid":"33c7560628f0fe446b6355d26f9fee67da4fc58a","id":"9WSL1H-WipGf","colab_type":"code","outputId":"4c195237-feda-45b7-c34d-7c83424afd60","executionInfo":{"status":"ok","timestamp":1554054263729,"user_tz":-330,"elapsed":1033629,"user":{"displayName":"Sup H","photoUrl":"","userId":"04337592286741421479"}},"colab":{"base_uri":"https://localhost:8080/","height":34}},"cell_type":"code","source":["test_datagen = ImageDataGenerator(rescale=1./255.)\n","\n","test_generator = test_datagen.flow_from_dataframe(dataframe=test_df, \n","                                                    directory=train_dir, \n","                                                    x_col=\"image_name\", \n","                                                    #y_col=\"label\", \n","                                                    class_mode=None,\n","                                                    target_size=train_input_shape[0:2], \n","                                                    batch_size=1,\n","                                                    shuffle=False\n","                                                 )"],"execution_count":0,"outputs":[{"output_type":"stream","text":["Found 7301 images.\n"],"name":"stdout"}]},{"metadata":{"_uuid":"08afdd8c3000a63d5846644d83bebd55bb2e7fdb","id":"6hixVyr-ipGw","colab_type":"text"},"cell_type":"markdown","source":["## Predict and Submit"]},{"metadata":{"_uuid":"067733fbfd186fa1d65d4110165b345a6c9f1898","id":"xX7gSEVUipG5","colab_type":"code","outputId":"d30ab836-82ac-4660-d70c-e35df2ea1abc","executionInfo":{"status":"ok","timestamp":1554054469678,"user_tz":-330,"elapsed":1236892,"user":{"displayName":"Sup H","photoUrl":"","userId":"04337592286741421479"}},"colab":{"base_uri":"https://localhost:8080/","height":51}},"cell_type":"code","source":["# Predict on test data\n","test_generator.reset()\n","predictions = model.predict_generator(test_generator,verbose=1)\n","predictions = np.argmax(predictions, axis=1)\n","#predictions = predictions.astype('int').ravel()\n","predictions.shape"],"execution_count":0,"outputs":[{"output_type":"stream","text":["7301/7301 [==============================] - 206s 28ms/step\n"],"name":"stdout"},{"output_type":"execute_result","data":{"text/plain":["(7301,)"]},"metadata":{"tags":[]},"execution_count":41}]},{"metadata":{"_uuid":"3c52db7624172d624771f0a00205568676fcc853","id":"QK9ubPZwipHB","colab_type":"code","outputId":"c1b214bb-8602-4dab-f2ae-a315b0a004ac","executionInfo":{"status":"ok","timestamp":1554054469682,"user_tz":-330,"elapsed":1236021,"user":{"displayName":"Sup H","photoUrl":"","userId":"04337592286741421479"}},"colab":{"base_uri":"https://localhost:8080/","height":34}},"cell_type":"code","source":["# Retrieve filenames\n","import re\n","#test_img_ids = [re.split(\"/\", val)[1] for val in test_generator.filenames]\n","test_img_ids = test_generator.filenames\n","len(test_img_ids)"],"execution_count":0,"outputs":[{"output_type":"execute_result","data":{"text/plain":["7301"]},"metadata":{"tags":[]},"execution_count":42}]},{"metadata":{"_uuid":"5785b0499ce1d7425bce81e98c2d20d29dc7cc7e","id":"QpsDW_baipHL","colab_type":"code","colab":{}},"cell_type":"code","source":["# Create dataframe for submission\n","submission_df = pd.DataFrame({'image_name' : test_img_ids,\n","                              'label' : predictions })\n","submission_df.head()"],"execution_count":0,"outputs":[]},{"metadata":{"_uuid":"f2b7c10b6dd4103cf2477fe6f492a635f312b1e1","id":"O16A0M70ipHW","colab_type":"code","colab":{}},"cell_type":"code","source":["# Create submission file\n","submission_dir = os.path.join(home_dir, \"submissions\")\n","submission_df.to_csv(submission_dir + '/submission_resnet50.csv', index=False)"],"execution_count":0,"outputs":[]}]}