{"cells":[{"metadata":{},"cell_type":"markdown","source":"# This notebook calculates shape and aggreagation landscape metrics for melanomas.... only 46 here but over 130 extras features are available..","execution_count":null},{"metadata":{},"cell_type":"markdown","source":"# it could capture additional tabular information to use to enhance predictions with NNs.","execution_count":null},{"metadata":{"_uuid":"c0bec335-dffe-4490-bc34-8441eadbdbad","_cell_guid":"0dd91320-9ae0-40d5-879b-553279541c8b","trusted":true},"cell_type":"code","source":"!conda install -c conda-forge -y pylandstats\nimport numpy as np\nfrom pylandstats import *\nimport matplotlib.pyplot as plt","execution_count":null,"outputs":[]},{"metadata":{"_uuid":"dc2c067f-2921-48de-8b4c-9e134bcc4a40","_cell_guid":"280479d5-7ead-4318-890a-1c57baee50f8","trusted":true},"cell_type":"code","source":"train = np.load(\"../input/siimisic-melanoma-resized-images/x_train_32.npy\", mmap_mode=\"r\")\ntest = np.load(\"../input/siimisic-melanoma-resized-images/x_test_32.npy\", mmap_mode=\"r\")","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"row=np.random.choice(range(train.shape[0]))\narray=train[row,:,:,:]\nch=2\nbarray=np.where(array[:,:,ch] > np.quantile(array[:,:,ch],0.15), 2, 1)\n\nplt.figure(figsize=(10,5)) \n\nplt.subplot(1, 2, 1)\nplt.imshow(array, cmap=plt.cm.binary)\nplt.axis('off')\nplt.subplot(1, 2, 2)\nplt.imshow(barray, cmap=plt.cm.binary)\nplt.axis('off')\nplt.subplots_adjust(wspace=-0.1, hspace=-0.1)\nplt.show()","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"parray=Landscape(barray, res=(1,1))\nLSMetrics=parray.compute_landscape_metrics_df()\nparray.plot_landscape()\nLSMetrics.head()","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"TrainFeatures=np.zeros((train.shape[0],46,3))\n\nfor j in range(train.shape[0]):\n  for ch in range(3):\n    array=train[j,:,:,ch]\n    barray=np.where(array > np.quantile(array,0.15), 2, 1)\n    parray=Landscape(barray, res=(1,1))\n    LSMetrics=parray.compute_landscape_metrics_df()\n    TrainFeatures[j,:,ch]=LSMetrics[0:].values\n\nTestFeatures=np.zeros((test.shape[0],46,3))\n\nfor j in range(test.shape[0]):\n  for ch in range(3):\n    array=test[j,:,:,ch]\n    barray=np.where(array > np.quantile(array,0.15), 2, 1)\n    parray=Landscape(barray, res=(1,1))\n    LSMetrics=parray.compute_landscape_metrics_df()\n    TestFeatures[j,:,ch]=LSMetrics[0:].values","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"np.save(\"trainFeatures.npy\", TrainFeatures)\nnp.save(\"testFeatures.npy\", TestFeatures)","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":4}