seedling-classification-cnn-tensorflow.ipynb

{"cells":[{"metadata":{"_cell_guid":"98e315a1-ed2a-4c5b-a64a-926bbe720c57","_uuid":"4615543cb3b499e044c100725f41373b8d1ac2dd","collapsed":true,"trusted":true},"cell_type":"code","source":"import numpy as np #supporting multi-dimensional arrays and matrices\nimport os #read or write a file\nimport cv2  \nimport pandas as pd #data manipulation and analysis\nfrom tqdm import tqdm # for  well-established ProgressBar\nfrom random import shuffle #only shuffles the array along the first axis of a multi-dimensional array. The order of sub-arrays is changed but their contents remains the same.\nLR = 1e-3\nMODEL_NAME = 'plantclassfication-{}-{}.model'.format(LR, '2conv-basic') # just so we remember which save","execution_count":1,"outputs":[]},{"metadata":{"_cell_guid":"354cd553-6ed3-4b0d-ab9a-735be98cc539","_uuid":"ec22dd9eaab76f2c7f78377e02ba3cc31defc7e5","collapsed":true,"trusted":true},"cell_type":"code","source":"data_dir = '../input/'\ntrain_dir = os.path.join(data_dir, 'train')\ntest_dir = os.path.join(data_dir, 'test')\nIMG_SIZE = 128","execution_count":2,"outputs":[]},{"metadata":{"_cell_guid":"3b598efd-e3e1-4c7d-9538-c8f909e6c312","_uuid":"d1beab1ada8462d965c9e9954d6cbc7a3ed3103e","scrolled":true,"trusted":true},"cell_type":"code","source":"#list of categories in array format \nCATEGORIES = ['Black-grass', 'Charlock', 'Cleavers', 'Common Chickweed', 'Common wheat', 'Fat Hen', 'Loose Silky-bent',\n              'Maize', 'Scentless Mayweed', 'Shepherds Purse', 'Small-flowered Cranesbill', 'Sugar beet']\nNUM_CATEGORIES = len(CATEGORIES)\nprint (NUM_CATEGORIES)","execution_count":3,"outputs":[]},{"metadata":{"_cell_guid":"10b1ad72-e7c2-4a9f-a2de-bdc8e67bc771","_uuid":"4a37499662eecffba206cf6d44d0436c00cdf6f7","collapsed":true,"trusted":true},"cell_type":"code","source":"''' function that accept plant category and return array format of the vlaue , one-hot array\n am sure there's better way to do this .......'''\n\ndef label_img(word_label):                       \n    if word_label == 'Black-grass': return [1,0,0,0,0,0,0,0,0,0,0,0]\n    elif word_label == 'Charlock': return [0,1,0,0,0,0,0,0,0,0,0,0]\n    elif word_label == 'Cleavers': return [0,0,1,0,0,0,0,0,0,0,0,0]\n    elif word_label == 'Common Chickweed': return [0,0,0,1,0,0,0,0,0,0,0,0]\n    elif word_label == 'Common wheat': return [0,0,0,0,1,0,0,0,0,0,0,0]\n    elif word_label == 'Fat Hen': return [0,0,0,0,0,1,0,0,0,0,0,0]\n    elif word_label == 'Loose Silky-bent': return [0,0,0,0,0,0,1,0,0,0,0,0]\n    elif word_label == 'Maize': return [0,0,0,0,0,0,0,1,0,0,0,0]\n    elif word_label == 'Scentless Mayweed': return [0,0,0,0,0,0,0,0,1,0,0,0]\n    elif word_label == 'Shepherds Purse': return [0,0,0,0,0,0,0,0,0,1,0,0]\n    elif word_label == 'Small-flowered Cranesbill': return [0,0,0,0,0,0,0,0,0,0,1,0]\n    elif word_label == 'Sugar beet': return [0,0,0,0,0,0,0,0,0,0,0,1] ","execution_count":4,"outputs":[]},{"metadata":{"_cell_guid":"f5b8416e-7a60-4703-837c-37fd2cdb0aca","_uuid":"abb20d6b06d91a0594bccfef61eaef197eafc973","collapsed":true,"trusted":true},"cell_type":"code","source":"\n'''function that will create train data , will go thought all the file do this \n----read the image in  grayscale mode ,resize it\n---change it to numpy arrays and  append it to dataframe train with it`s associated category '''\n\ndef create_train_data():\n    train = []\n    for category_id, category in enumerate(CATEGORIES):\n        for img in tqdm(os.listdir(os.path.join(train_dir, category))):\n            label=label_img(category)\n            path=os.path.join(train_dir,category,img)\n            img=cv2.imread(path,cv2.IMREAD_GRAYSCALE)\n            img = cv2.resize(img, (IMG_SIZE,IMG_SIZE))\n            train.append([np.array(img),np.array(label)])\n    shuffle(train)\n    return train","execution_count":5,"outputs":[]},{"metadata":{"_cell_guid":"6ffc33b4-4deb-4907-9fd2-80fc4472e7e1","_uuid":"c9c563a0c5430ba3e482d680c22dbfb0105731e1","scrolled":true,"trusted":true},"cell_type":"code","source":"train_data = create_train_data()","execution_count":6,"outputs":[]},{"metadata":{"_cell_guid":"e6ca8c73-743a-4489-bbba-78a5482adb0c","_uuid":"e5937fcc03a8026d2203db89960a283161a98684","collapsed":true,"trusted":true},"cell_type":"code","source":"\n'''function that will create test data , will go thought  file do this \n----read the image in  grayscale mode ,resize it\n---change it to numpy arrays and  append it to dataframe test but no category here of course  '''\n\ndef create_test_data():\n    test = []\n    for img in tqdm(os.listdir(test_dir)):\n        path = os.path.join(test_dir,img)\n        img_num = img\n        img = cv2.imread(path,cv2.IMREAD_GRAYSCALE)\n        img = cv2.resize(img, (IMG_SIZE,IMG_SIZE))\n        test.append([np.array(img), img_num])\n        \n    shuffle(test)\n    return test","execution_count":7,"outputs":[]},{"metadata":{"_cell_guid":"c3b32ce1-8601-4b5f-86a6-d1c4d94993d5","_uuid":"eb9908cbe83e785fb569f0989bd1713f01349ccd","trusted":true},"cell_type":"code","source":"test_data = create_test_data()","execution_count":8,"outputs":[]},{"metadata":{"_cell_guid":"a1c2c83a-408d-4bf4-a947-a92cc187b22b","_uuid":"e36273911baf088bf396fae857074c7ac95f2f17","trusted":true},"cell_type":"code","source":"import tensorflow as tf #used for machine learning applications such as neural networks\nimport tflearn #modular and transparent deep learning library built on top of Tensorflow\nfrom tflearn.layers.conv import conv_2d, max_pool_2d \nfrom tflearn.layers.core import input_data, dropout, fully_connected\nfrom tflearn.layers.estimator import regression\ntf.reset_default_graph()\n\nconvnet = input_data(shape=[None, IMG_SIZE, IMG_SIZE, 1], name='input')\n\nconvnet = conv_2d(convnet, 32, 5, activation='relu')\nconvnet = max_pool_2d(convnet, 5)\n\nconvnet = conv_2d(convnet, 64, 5, activation='relu')\nconvnet = max_pool_2d(convnet, 5)\n\nconvnet = conv_2d(convnet, 32, 5, activation='relu')\nconvnet = max_pool_2d(convnet, 5)\n\nconvnet = conv_2d(convnet, 64, 5, activation='relu')\nconvnet = max_pool_2d(convnet, 5)\n\nconvnet = conv_2d(convnet, 32, 5, activation='relu')\nconvnet = max_pool_2d(convnet, 5)\n\nconvnet = conv_2d(convnet, 64, 5, activation='relu')\nconvnet = max_pool_2d(convnet, 5)\n\nconvnet = fully_connected(convnet, 1024, activation='relu')\nconvnet = dropout(convnet, 0.8)\n\nconvnet = fully_connected(convnet, 12, activation='softmax')\nconvnet = regression(convnet, optimizer='adam', learning_rate=LR, loss='categorical_crossentropy', name='targets')\n\nmodel = tflearn.DNN(convnet, tensorboard_dir='log')\n\n\n\nif os.path.exists('{}.meta'.format(MODEL_NAME)):\n    model.load(MODEL_NAME)\n    print('model loaded!')\n\ntrain = train_data\ntest = train_data\n\nX = np.array([i[0] for i in train]).reshape(-1,IMG_SIZE,IMG_SIZE,1)\nY = [i[1] for i in train]\n\ntest_x = np.array([i[0] for i in test]).reshape(-1,IMG_SIZE,IMG_SIZE,1)\ntest_y = [i[1] for i in test]\n\nmodel.fit({'input': X}, {'targets': Y}, n_epoch=5, validation_set=({'input': test_x}, {'targets': test_y}), \n    snapshot_step=500, show_metric=True, run_id=MODEL_NAME)","execution_count":9,"outputs":[]},{"metadata":{"_cell_guid":"da103282-4011-4367-b947-dc8dc9b02d37","_uuid":"65e81dd00abf2d6dd88324b86d4721e3ef0521e4","trusted":true},"cell_type":"code","source":"model.save(MODEL_NAME)","execution_count":10,"outputs":[]},{"metadata":{"_cell_guid":"3c9cffd6-9740-46b4-a4b8-dd97dd4c3bc7","_uuid":"b185638d2ead940493511ddcc2e307b3c28cfd9a","collapsed":true,"trusted":true},"cell_type":"code","source":"#return Indexes of the maximal elements of a array\ndef label_return (model_out):\n    if np.argmax(model_out) == 0: return  'Black-grass'\n    elif np.argmax(model_out) == 1: return 'Charlock'\n    elif np.argmax(model_out) == 2: return 'Cleavers'\n    elif np.argmax(model_out) == 3: return 'Common Chickweed'\n    elif np.argmax(model_out) == 4: return 'Common wheat'\n    elif np.argmax(model_out) == 5: return 'Fat Hen'\n    elif np.argmax(model_out) == 6: return 'Loose Silky-bent'\n    elif np.argmax(model_out) == 7: return 'Maize'\n    elif np.argmax(model_out) == 8: return 'Scentless Mayweed'\n    elif np.argmax(model_out) == 9: return 'Shepherds Purse'\n    elif np.argmax(model_out) == 10: return 'Small-flowered Cranesbill'\n    elif np.argmax(model_out) == 11: return 'Sugar beet'\n    ","execution_count":11,"outputs":[]},{"metadata":{"_cell_guid":"70866b22-ba31-4ea4-96be-43c24e3e76ad","_uuid":"b2fd863bf2fa70338d9133ccead6d3ff7ce11c16","trusted":true},"cell_type":"code","source":"import matplotlib.pyplot as plt\n\ntest_data = create_test_data()\nfig=plt.figure(figsize = (18,10))\nfor num,data in enumerate(test_data[:12]): \n    img_num = data[1]\n    img_data = data[0]\n    y = fig.add_subplot(3,4,num+1)\n    orig = img_data\n    data = img_data.reshape(IMG_SIZE,IMG_SIZE,1)\n    model_out = model.predict([data])[0]\n    str_label=label_return (model_out)\n    y.imshow(orig,cmap='gray',interpolation='nearest')\n    plt.title(str_label)\n    y.axes.get_xaxis().set_visible(False)\n    y.axes.get_yaxis().set_visible(False)\nplt.show()\n","execution_count":16,"outputs":[]},{"metadata":{"_cell_guid":"fa395220-117e-4505-8bfc-3b0fca42e1ef","_uuid":"99f25e27a8522b13148b9cb5e2d4b4e2f1ad6d9b","trusted":true},"cell_type":"code","source":"import pandas as pd\nsample_submission = pd.read_csv('../input/sample_submission.csv')\nsample_submission.head(2)","execution_count":13,"outputs":[]},{"metadata":{"_cell_guid":"ed337054-c982-4d63-95d5-12586f64f4a5","_uuid":"45ce884049ce770a3d1dc963e7dc0b37c8ccdfe3","trusted":true},"cell_type":"code","source":"test_data = create_test_data()\nwith open('sample_submission.csv','w') as f:\n    f.write('file,species\\n')\n    for data in test_data:\n        img_num = data[1]\n        img_data = data[0]\n        orig = img_data\n        data = img_data.reshape(IMG_SIZE,IMG_SIZE,1)\n        model_out = model.predict([data])[0]\n        str_label=label_return (model_out)\n        file = img_num\n        species = str_label\n        row = file + \",\" + species + \"\\n\"\n        f.write(row)","execution_count":14,"outputs":[]},{"metadata":{"_cell_guid":"9f4b43bd-98d9-4229-8892-e74f5e64dccd","_uuid":"b2b3259dfcc9f5a857c7adb8a233e7574e5351b4","scrolled":true,"trusted":true},"cell_type":"code","source":"import pandas as pd\nsample_submission = pd.read_csv('sample_submission.csv')\nsample_submission.head(10)","execution_count":null,"outputs":[]},{"metadata":{"_cell_guid":"daf1fd37-803e-46fa-8d05-f4351b604efb","_uuid":"312cad0b8a134785c8a9d7271ea27773d82d4d9a","collapsed":true,"trusted":true},"cell_type":"code","source":"","execution_count":null,"outputs":[]}],"metadata":{"kernelspec":{"display_name":"Python 3","language":"python","name":"python3"},"language_info":{"name":"python","version":"3.6.4","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"}},"nbformat":4,"nbformat_minor":1}