{"cells":[{"metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true,"collapsed":true},"cell_type":"code","source":"from skimage.morphology import closing, disk, label\nfrom scipy.ndimage.morphology import binary_closing\nfrom PIL import Image, ImageFilter\nfrom multiprocessing import *\nimport matplotlib.pyplot as plt\n%matplotlib inline\nimport pandas as pd\nimport numpy as np\nimport cv2\n\npath = '../input/'\ntrain = pd.read_csv('../input/train_ship_segmentations.csv')\ntest = pd.read_csv('../input/sample_submission.csv')\nprint(train.shape, test.shape)","execution_count":null,"outputs":[]},{"metadata":{"_cell_guid":"79c7e3d0-c299-4dcb-8224-4455121ee9b0","_uuid":"d629ff2d2480ee46fbb7e2d37f6b5fab8052498a","trusted":true,"collapsed":true},"cell_type":"code","source":"img = np.random.choice(train.ImageId.values)\nim = Image.open(path + 'train/' + img).convert('RGB')\nplt.imshow(im)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"5e8f32ad126c3d034c2f13a94b4f06ce71e53e74","collapsed":true},"cell_type":"code","source":"im = im.filter(ImageFilter.EMBOSS).convert('L')\nim = (np.array(im)> 150).astype(np.uint8)\nim = closing(im, selem=disk(5))\nplt.imshow(im)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"8158ef3b3f9a004a7de409583e29253a0e7dd5c5","collapsed":true},"cell_type":"code","source":"%%time\n#https://www.kaggle.com/rakhlin/fast-run-length-encoding-python\ndef rl_encoding(img):\n    im = Image.open('../input/test/' + img).convert('RGB')\n    im = im.filter(ImageFilter.EMBOSS).convert('L')\n    im = (np.array(im)> 150).astype(np.uint8)\n    im = closing(im, selem=disk(5))\n    dots = np.where(im.T.flatten()==1)[0]\n    run_lengths = []\n    prev = -2\n    for b in dots:\n        if (b>prev+1): run_lengths.extend((b+1, 0))\n        run_lengths[-1] += 1\n        prev = b\n    run_lengths = np.array(run_lengths).reshape(len(run_lengths)//2,2)\n    run_lengths = [' '.join(map(str, [x, y])) for x, y in run_lengths if y > 2] #limit\n    run_lengths = ' '.join(map(str, run_lengths))\n    return [img, run_lengths]\n\np = Pool(cpu_count())\nresults = p.map(rl_encoding, test['ImageId'].values[:8000])\np.close(); p.join()","execution_count":null,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"218834bf57079f4f6fbe74fd2bbb4237e6220f32","collapsed":true},"cell_type":"code","source":"sub = pd.DataFrame(results)\nsub.columns = ['ImageId', 'EncodedPixels']\nsub1 = pd.read_csv('../input/sample_submission.csv')\nsub1 = pd.DataFrame(np.setdiff1d(sub1['ImageId'].unique(), sub['ImageId'].unique(), assume_unique=True), columns=['ImageId'])\nsub1['EncodedPixels'] = None #'1 2'\nprint(len(sub1), len(sub))\nsub = pd.concat([sub, sub1])\nprint(len(sub))\nsub.to_csv('submission.csv', index=False)","execution_count":null,"outputs":[]}],"metadata":{"kernelspec":{"display_name":"Python 3","language":"python","name":"python3"},"language_info":{"name":"python","version":"3.6.6","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"}},"nbformat":4,"nbformat_minor":1}