{"cells":[{"metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true},"cell_type":"code","source":"# This Python 3 environment comes with many helpful analytics libraries installed\n# It is defined by the kaggle/python docker image: https://github.com/kaggle/docker-python\n# For example, here's several helpful packages to load in \n\nimport numpy as np # linear algebra\nimport pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\n\n# Input data files are available in the \"../input/\" directory.\n# For example, running this (by clicking run or pressing Shift+Enter) will list all files under the input directory\n\nimport os\n#for dirname, _, filenames in os.walk('/kaggle/input'):\n#    for filename in filenames:\n#        print(os.path.join(dirname, filename))\nprint(os.listdir(\"../input\"))\n# Any results you write to the current directory are saved as output.","execution_count":null,"outputs":[]},{"metadata":{"_uuid":"d629ff2d2480ee46fbb7e2d37f6b5fab8052498a","collapsed":true,"_cell_guid":"79c7e3d0-c299-4dcb-8224-4455121ee9b0","trusted":false},"cell_type":"code","source":"import keras\nimport numpy as np\nimport pandas as pd\nfrom PIL import Image as img\nimport glob\nfrom tqdm import tqdm\nimport matplotlib.pyplot as plt\nfrom keras.models import Sequential\nfrom keras.layers import Dense, Dropout, Flatten\nfrom keras.layers import Conv2D, MaxPooling2D\nfrom sklearn.model_selection import train_test_split\nfrom keras.utils import to_categorical\nfrom keras.preprocessing import image\nkeras.backend.image_data_format()\nkeras.backend.set_image_data_format(\"channels_first\")\nkeras.backend.image_data_format()\n%matplotlib inline","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"y_train = pd.read_csv(\"/kaggle/input/aptos2019-blindness-detection/train.csv\")\nblind_train_list = []\nfor i in list(y_train['id_code']):\n    path = \"/kaggle/input/aptos2019-blindness-detection/train_images/\"+i+\".png\"\n    blind_train_list.append(path)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"x_train = []\nfor i in tqdm(blind_train_list):\n    temp = img.open(i).resize((64, 64))\n    temp = temp.convert(\"L\") \n    x_train.append((np.array(temp) - np.mean(temp)) / np.std(temp))\nprint(\"aptos blindness detection images loading is done\")\n\na = np.asarray(x_train)\nx_train = a.reshape(a.shape[0], 1, a.shape[1], a.shape[2])","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"y_train_ = y_train['diagnosis']\ny_train_df = pd.DataFrame()\nzero = []\none = []\ntwo = []\nthree = []\nfour = []\nfor i in list(y_train_):\n    if i == 0:\n        zero.append(1)\n        one.append(0)\n        two.append(0)\n        three.append(0)\n        four.append(0)\n    elif i == 1:\n        zero.append(0)\n        one.append(1)\n        two.append(0)\n        three.append(0)\n        four.append(0)\n    elif i == 2:\n        zero.append(0)\n        one.append(0)\n        two.append(1)\n        three.append(0)\n        four.append(0)\n    elif i == 3:\n        zero.append(0)\n        one.append(0)\n        two.append(1)\n        three.append(0)\n        four.append(0)\n    elif i == 4:\n        zero.append(0)\n        one.append(0)\n        two.append(0)\n        three.append(0)\n        four.append(1)\n\ny_train_df = pd.DataFrame({'zero':zero,'one':one,'two':two,'three':three,'four':four})  \ny_train_val = np.array(y_train_df)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"model = Sequential()\nmodel.add(Conv2D(filters=16, kernel_size=(5, 5), activation=\"relu\", input_shape=(1,64,64)))\nmodel.add(MaxPooling2D(pool_size=(2, 2)))\nmodel.add(Dropout(0.25))\nmodel.add(Conv2D(filters=32, kernel_size=(5, 5), activation='relu'))\nmodel.add(MaxPooling2D(pool_size=(2, 2)))\nmodel.add(Dropout(0.25))\nmodel.add(Conv2D(filters=64, kernel_size=(5, 5), activation='relu'))\nmodel.add(MaxPooling2D(pool_size=(2, 2)))\nmodel.add(Dropout(0.25))\nmodel.add(Flatten())\nmodel.add(Dense(128, activation='relu'))\nmodel.add(Dropout(0.5))\nmodel.add(Dense(64, activation='relu'))\nmodel.add(Dropout(0.5))\nmodel.add(Dense(5, activation='softmax'))\n\nmodel.compile(optimizer='adam', loss='binary_crossentropy', metrics=['accuracy'])\nmodel.summary()","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"X_train_, X_val, y_train_, y_val = train_test_split(x_train, y_train_val, random_state=42, test_size=0.1)\nX_train, X_test, y_train, y_test = train_test_split(X_train_, y_train_, random_state=42, test_size=0.1)\nmodel.fit(X_train, y_train, epochs=30, validation_data=(X_test, y_test), batch_size=100)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"plt.figure(figsize=(20, 7))\nplt.subplot(1, 2, 1)\nplt.plot(model.history.history[\"acc\"])\nplt.plot(model.history.history[\"val_acc\"])\nplt.title(\"Model Accuracy\")\nplt.ylabel(\"Accuracy\")\nplt.xlabel(\"Epoch\")\nplt.legend([\"Train\", \"Val\"], loc=\"upper left\")\n\nplt.xticks(np.arange(len(model.history.history[\"acc\"])), np.arange(1, len(model.history.history[\"val_acc\"])+1, 1))\n\nplt.subplot(1, 2, 2)\nplt.plot(model.history.history[\"loss\"])\nplt.plot(model.history.history[\"val_loss\"])\nplt.title(\"Model Loss\")\nplt.ylabel(\"Loss\")\nplt.xlabel(\"Epoch\")\nplt.legend([\"Train\", \"Val\"], loc=\"upper right\")\nplt.xticks(np.arange(len(model.history.history[\"loss\"])), np.arange(1, len(model.history.history[\"loss\"])+1, 1))\nplt.show()","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"result_val = model.predict(x=X_val)\ndf_val = pd.DataFrame(result_val)\npred_val_list = []\nfor j in range(367):\n    zero_ = float(df_val.iloc[j,0:1])\n    one_ = float(df_val.iloc[j,1:2])\n    two_ = float(df_val.iloc[j,2:3])\n    thre_ = float(df_val.iloc[j,3:4])\n    four_ = float(df_val.iloc[j,4:])\n    \n    if zero_ > one_ and zero_ > two_ and zero_ > thre_ and zero_ > four_:\n        pred_val_list.append(0)\n    elif one_ > zero_ and one_ > two_ and one_ > thre_ and one_ > four_:\n        pred_val_list.append(1)\n    elif two_ > zero_ and two_ > one_ and two_ > thre_ and two_ > four_:\n        pred_val_list.append(2)\n    elif thre_ > zero_ and thre_ > one_ and thre_ > two_ and thre_ > four_:\n        pred_val_list.append(3)\n    elif four_ > zero_ and four_ > one_ and four_ > two_ and four_ > thre_:\n        pred_val_list.append(4)\n    \ndf_val_ = pd.DataFrame(y_val)\nval_list = []\nfor j in range(367):\n    zero_ = float(df_val_.iloc[j,0:1])\n    one_ = float(df_val_.iloc[j,1:2])\n    two_ = float(df_val_.iloc[j,2:3])\n    thre_ = float(df_val_.iloc[j,3:4])\n    four_ = float(df_val_.iloc[j,4:])\n    \n    if zero_ > one_ and zero_ > two_ and zero_ > thre_ and zero_ > four_:\n        val_list.append(0)\n    elif one_ > zero_ and one_ > two_ and one_ > thre_ and one_ > four_:\n        val_list.append(1)\n    elif two_ > zero_ and two_ > one_ and two_ > thre_ and two_ > four_:\n        val_list.append(2)\n    elif thre_ > zero_ and thre_ > one_ and thre_ > two_ and thre_ > four_:\n        val_list.append(3)\n    elif four_ > zero_ and four_ > one_ and four_ > two_ and four_ > thre_:\n        val_list.append(4)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"df_comb = pd.DataFrame({'actual':val_list,'predicted':pred_val_list})\nsm = 0\nfor i in range(367):\n    fst = int(df_comb.iloc[i,0:1])\n    snd =  int(df_comb.iloc[i,1:])\n    if fst == snd:\n        sm = sm + 1  \nsm/df_comb.shape[0]","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"y_test = pd.read_csv(\"/kaggle/input/aptos2019-blindness-detection/test.csv\")\nblind_test_list = []\nfor i in list(y_test['id_code']):\n    path = \"/kaggle/input/aptos2019-blindness-detection/test_images/\"+i+\".png\"\n    blind_test_list.append(path)\n    \nx_test = []\nfor i in tqdm(blind_test_list):\n    temp_test = img.open(i).resize((64, 64))\n    temp_test = temp_test.convert(\"L\")\n    x_test.append((np.array(temp_test) - np.mean(temp_test)) / np.std(temp_test))\nprint(\"aptos blindness test images loading is done\")\nb = np.asarray(x_test)\nx_test = b.reshape(b.shape[0], 1, b.shape[1], b.shape[2])","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"result_test = model.predict(x=x_test)\ndf_test = pd.DataFrame(result_test)\npred_test_list = []\nfor j in range(1928):\n    zero_ = float(df_test.iloc[j,0:1])\n    one_ = float(df_test.iloc[j,1:2])\n    two_ = float(df_test.iloc[j,2:3])\n    thre_ = float(df_test.iloc[j,3:4])\n    four_ = float(df_test.iloc[j,4:])\n    \n    if zero_ > one_ and zero_ > two_ and zero_ > thre_ and zero_ > four_:\n        pred_test_list.append(0)\n    elif one_ > zero_ and one_ > two_ and one_ > thre_ and one_ > four_:\n        pred_test_list.append(1)\n    elif two_ > zero_ and two_ > one_ and two_ > thre_ and two_ > four_:\n        pred_test_list.append(2)\n    elif thre_ > zero_ and thre_ > one_ and thre_ > two_ and thre_ > four_:\n        pred_test_list.append(3)\n    elif four_ > zero_ and four_ > one_ and four_ > two_ and four_ > thre_:\n        pred_test_list.append(4)\n        \ny_test_ = y_test['id_code']\nsub = pd.DataFrame({'id_code':y_test_,'diagnosis':pred_test_list})\nsub.to_csv(\"../input/sampleSubmission.csv\",index=False)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"pd.read_csv(\"../input/sampleSubmission.csv\").to_csv('../sample_Submission.csv')\n#print(sub_df.head(5))","execution_count":null,"outputs":[]}],"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"}},"nbformat":4,"nbformat_minor":1}