{"cells":[{"metadata":{},"cell_type":"markdown","source":"# Prostate cANcer graDe Assessment (PANDA) Challenge: Creating the Training Dataset"},{"metadata":{},"cell_type":"markdown","source":"# Introduction"},{"metadata":{},"cell_type":"markdown","source":"## Prostate Gland and Prostate Cancer"},{"metadata":{},"cell_type":"markdown","source":"The prostate is part of the male reproductive system.  The function of the prostate gland is to secrete substances to the urethra.  These secretions nurish and transport sperm.  Prostate cancer is diagnosed from samples from a prostate biopsy (rmicrobe).  The sample is first assigned as a gleason score.  This score is converted to a ISUP grade of 0-5.  The score of 0 is negative and the score of 5 is the most severe form of cancer (Prostate cANcer GraDe Assessment (PANDA) Challenge).\n\nThe purpose of this notebook is to use fast.ai and ResNet34 to predict the ISUP grade of different prostate cancer biopsies."},{"metadata":{},"cell_type":"markdown","source":"## Current Problems With Image Processing and Training"},{"metadata":{},"cell_type":"markdown","source":"<font color='red'>Problem 1:</font> The biopsy has a large size and it is difficult to observe the glands in the biopsy.\n\n<font color='green'>Solution 1:</font> Randomly generate patches of the biopsy to magnify the stroma of the prostate.\n\n<font color='red'>Problem 2:</font> The randomly generated patches may only show blank areas and the prostate capsule.\n\n<font color='green'>Solution 2:</font> Remove gray areas around the biopsy and generate patches that are generated from areas that have a minimal amount of white area.  This increases the probability that the patches will be in the prostate stroma.\n\n<font color='red'>Problem 3:</font> Despite removing suspicious slides, there are slides with blue ink and slides that appear to be stained improperly.  Some images may also be difficult to discern because of the glandular pattern or the patch generation.\n\n<font color='green'>Solution 3:</font> Removing high error biopsy slides would ensure that the training images are appropriate for creating a training model."},{"metadata":{},"cell_type":"markdown","source":"## Image Processing and Creating the Dataset "},{"metadata":{},"cell_type":"markdown","source":"1. Remove the gray area surrounding the biopsy.  The first step involves removing the gray area from around the biopsy (Zenify). \n2. Create 4X4 patched image.  The second step is to take 16 samples that have the lowest portion of white.  This ensures that the sample is most likely going to show the appropriate part of the sample (i.e. glands). (PAB97).\n3. Use fast.ai and ResNet34 to train and predict ISUP grades.\n4. Remove High Error Biopsies.  Remove biopsies with a high error difference.  The error difference will be at least two ISUP grades higher than the assigned ISUP.\n5. Use fast.ai and ResNet34 to train and predict ISUP grades.\n6. Remove High Error Biopsies.  Remove biopsies with a high error difference.  The error difference will be at least two ISUP grades higher than the assigned ISUP.\n7. This will be the dataset for training."},{"metadata":{},"cell_type":"markdown","source":"![image.png](attachment:image.png)","attachments":{"image.png":{"image/png":"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"}}},{"metadata":{},"cell_type":"markdown","source":"# Import fast.ai and Dependencies"},{"metadata":{},"cell_type":"markdown","source":"## Install fastai without Internet"},{"metadata":{},"cell_type":"markdown","source":"Internet is not allowed in this competition.  The files have to be loaded through the fastai2 dataset."},{"metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true},"cell_type":"code","source":"!pip install ../input/fastai2/fastprogress-0.2.3-py3-none-any.whl\n!pip install ../input/fastai2/fastcore-0.1.18-py3-none-any.whl\n!pip install ../input/fastai2/fastai2-0.0.17-py3-none-any.whl","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"## Import fast.ai"},{"metadata":{"trusted":true},"cell_type":"code","source":"from fastai2.basics import *\nfrom fastai2.callback.all import *\nfrom fastai2.vision.all import *","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"## Load Dependencies"},{"metadata":{"trusted":true},"cell_type":"code","source":"import os\nimport cv2\nimport PIL\nfrom PIL import Image as Img\nfrom PIL import ImageTk\nimport random\nimport openslide\nimport skimage.io\nimport skimage.color\nfrom skimage.color import rgb2hsv\nimport matplotlib\nimport numpy as np\nimport pandas as pd\nimport seaborn as sns\nimport matplotlib.pyplot as plt\nfrom IPython.display import Image, display","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Set Random Seed"},{"metadata":{},"cell_type":"markdown","source":"Setting a random seed makes sure that all randomly picked sequences are in the same order.  It is important to keep the same seed everytime to keep the same sequence everytime."},{"metadata":{"_uuid":"d629ff2d2480ee46fbb7e2d37f6b5fab8052498a","_cell_guid":"79c7e3d0-c299-4dcb-8224-4455121ee9b0","trusted":true},"cell_type":"code","source":"np.random.seed(2)","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Import the Prostate Cancer Grade Assessment Dataframe"},{"metadata":{},"cell_type":"markdown","source":"Import the PANDAS .csv file from the PANDAS directory.  This will be the PANDAS dataframe.  The other file came from [panda-analysis](http://https://www.kaggle.com/dannellyz/panda-analysis).  This database included suspicious slides that either had no masks or marking on them."},{"metadata":{"trusted":true},"cell_type":"code","source":"train_df = pd.read_csv('../input/prostate-cancer-grade-assessment/train.csv', index_col=False)\n\nsuspicious_slides = pd.read_csv('../input/panda-analysis/PANDA_Suspicious_Slides.csv', index_col=False)\n\ndata_dir = '../input/prostate-cancer-grade-assessment/train_images/'","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"It is important to get a preview of the suspicious_slides and the train_df dataframes."},{"metadata":{"trusted":true},"cell_type":"code","source":"suspicious_slides.head()","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"train_df.head()","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"Some suspicious slides filenames have been removed from train.csv.  These rows must be removed from the suspicious_slides dataframe before removing the suspicious slides from train.csv.  If they are not removed, an error will occur."},{"metadata":{"trusted":true},"cell_type":"code","source":"suspicious_slides = suspicious_slides.set_index('image_id')\n\nsuspicious_slides.drop (['01dfcde514052a6dc35ea4407f41d6e1'], inplace = True)\n\nsuspicious_slides.drop (['501403e5c13c7b86bc9312d087a6e490'], inplace = True)\n\nsuspicious_slides.drop (['8be88cbd606502cf980f721d57c1c794'], inplace = True)\n\nsuspicious_slides.drop (['9665b32251dfc1c438787d36e8b66dd0'], inplace = True)\n\nsuspicious_slides.drop (['c0852776d3142ce132a90c6142bd7146'], inplace = True)\n\nsuspicious_slides.drop (['cdb7663719497428b4d9243b76da3ace'], inplace = True)\n\nsuspicious_slides.drop (['d6b1c8ca6037b5ddca5d2086975a643b'], inplace = True)\n\nsuspicious_slides.drop (['3790f55cad63053e956fb73027179707'], inplace = True)\n\nsuspicious_slides.drop (['4a2ca53f240932e46eaf8959cb3f490a'], inplace = True)\n\nsuspicious_slides.drop (['e4215cfc8c41ec04a55431cc413688a9'], inplace = True)\n\nsuspicious_slides.drop (['aaa5732cd49bffddf0d2b7d36fbb0a83'], inplace = True)\n\nsuspicious_slides = suspicious_slides.reset_index()","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"markdown","source":"Delete the suspicous slides before the analysis."},{"metadata":{"trusted":true},"cell_type":"code","source":"suspicious = list (suspicious_slides['image_id'])\n\ntrain_df = train_df.set_index('image_id')\n\nfor i in suspicious:\n    train_df.drop([i], inplace=True)","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Exploratory Data Analysis of the PANDAS Dataframe"},{"metadata":{},"cell_type":"markdown","source":"It is important to get basic statistics on the PANDAS dataframe to understand the distribution of the data and to make sure the data is labeled properly."},{"metadata":{"trusted":true},"cell_type":"code","source":"train_df.head ()","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"train_df = train_df.reset_index()\n\ntrain_df.head()","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"**Get the number of IDs (biopsy samples) for the dataframe.**"},{"metadata":{"trusted":true},"cell_type":"code","source":"print(\"Number of IDs: \", len(train_df.image_id.unique()))","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"**Make sure the Data Providers and ISUP Grades are labeled properly for the dataframe.**"},{"metadata":{"trusted":true},"cell_type":"code","source":"print(\"Number of Data Providers: \", len(train_df.data_provider.unique()))\nprint(\"Number of ISUP Grades: \", len(train_df.isup_grade.unique()))","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"**Make sure the Gleason Scores are labeled properly for the dataframe and has the proper ISUP Grade label.**"},{"metadata":{"trusted":true},"cell_type":"code","source":"print(\"Number of Gleason Scores: \", len(train_df.gleason_score.unique()))","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"train_df['gleason_score'].unique()","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"Check the negative labels of the dataframe and make sure it has the proper ISUP Grade label."},{"metadata":{"trusted":true},"cell_type":"code","source":"print(train_df[train_df['gleason_score']=='0+0']['isup_grade'].unique())\nprint(train_df[train_df['gleason_score']=='negative']['isup_grade'].unique())","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"Check to make sure the 3+4, 4+3, 3+5, 5+3, 5+4, and 4+5 gleason score are labeled properly and make sure it has the proper ISUP Grade label."},{"metadata":{"trusted":true},"cell_type":"code","source":"print(train_df[(train_df['gleason_score']=='3+4') | (train_df['gleason_score']=='4+3')]['isup_grade'].unique())\nprint(train_df[(train_df['gleason_score']=='3+5') | (train_df['gleason_score']=='5+3')]['isup_grade'].unique())\nprint(train_df[(train_df['gleason_score']=='5+4') | (train_df['gleason_score']=='4+5')]['isup_grade'].unique())","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"Check to see if the Gleason Score 3+4 or 4+3 may have the improper ISUP Grade label."},{"metadata":{"trusted":true},"cell_type":"code","source":"print(train_df[train_df['gleason_score']=='3+4']['isup_grade'].unique())\nprint(train_df[train_df['gleason_score']=='4+3']['isup_grade'].unique())","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"Find the mislabeled sample for 4+3 and get rid of the ID."},{"metadata":{"trusted":true},"cell_type":"code","source":"train_df[(train_df['isup_grade'] == 2) & (train_df['gleason_score'] == '4+3')]","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"train_df.drop([6852],inplace=True)","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"**Recheck the numbers.**"},{"metadata":{"trusted":true},"cell_type":"code","source":"print(\"Number of IDs: \", len(train_df.image_id.unique()))\nprint(\"Number of Data Providers: \", len(train_df.data_provider.unique()))\nprint(\"Number of ISUP Grades: \", len(train_df.isup_grade.unique()))\nprint(\"Number of Gleason Scores: \", len(train_df.gleason_score.unique()))","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Import the Prostate Train Dataframe"},{"metadata":{},"cell_type":"markdown","source":"Import the four .csv file from the prostate train dataset directory. This will be the training dataframe for this part of the process.  The code will be commented out of for this version of the notebook.  To use this code, comment out the code under the **Import the Prostate Cancer Grade Assessment Dataframe** and **Exploratory Data Analysis of the PANDAS Dataframe**."},{"metadata":{"trusted":true},"cell_type":"code","source":"#train_df1 = pd.read_csv('../input/prostate-train-dataset/prediction_df1.csv', index_col=False)\n\n#train_df2 = pd.read_csv('../input/prostate-train-dataset/prediction_df2.csv', index_col=False)\n\n#train_df3 = pd.read_csv('../input/prostate-train-dataset/prediction_df3.csv', index_col=False)\n\n#train_df4 = pd.read_csv('../input/prostate-train-dataset/prediction_df4.csv', index_col=False)\n\n#data_dir = '../input/prostate-cancer-grade-assessment/train_images/'","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"Append the four dataframes and get rid of the 'Unnamed:0' column."},{"metadata":{"trusted":true},"cell_type":"code","source":"#train_df = train_df1.append(train_df2, ignore_index=True)\n#train_df = train_df.append(train_df3, ignore_index=True)\n#train_df = train_df.append(train_df4, ignore_index=True)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"#train_df = train_df.drop(columns=['Unnamed: 0'])","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"Get a preview of the dataframe."},{"metadata":{"trusted":true},"cell_type":"code","source":"#train_df.head ()","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"Reset the index to the dataframe."},{"metadata":{"trusted":true},"cell_type":"code","source":"#train_df = train_df.reset_index()\n\n#train_df.head()","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"It is important to get basic statistics on the PANDAS dataframe to understand the distribution of the data and to make sure the data is labeled properly."},{"metadata":{},"cell_type":"markdown","source":"**Get the number of IDs (biopsy samples) for the dataframe.**"},{"metadata":{"trusted":true},"cell_type":"code","source":"#print(\"Number of IDs: \", len(train_df.image_id.unique()))","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"**Make sure the Data Providers and ISUP Grades are labeled properly for the dataframe.**"},{"metadata":{"trusted":true},"cell_type":"code","source":"#print(\"Number of Data Providers: \", len(train_df.data_provider.unique()))\n#print(\"Number of ISUP Grades: \", len(train_df.isup_grade.unique()))","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"**Make sure the Gleason Scores are labeled properly for the dataframe and has the proper ISUP Grade label.**"},{"metadata":{"trusted":true},"cell_type":"code","source":"#print(\"Number of Gleason Scores: \", len(train_df.gleason_score.unique()))","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"#train_df['gleason_score'].unique()","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"Check the negative labels of the dataframe and make sure it has the proper ISUP Grade label."},{"metadata":{"trusted":true},"cell_type":"code","source":"#print(train_df[train_df['gleason_score']=='0+0']['isup_grade'].unique())\n#print(train_df[train_df['gleason_score']=='negative']['isup_grade'].unique())","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"Check to make sure the 3+4, 4+3, 3+5, 5+3, 5+4, and 4+5 gleason score are labeled properly and make sure it has the proper ISUP Grade label."},{"metadata":{"trusted":true},"cell_type":"code","source":"#print(train_df[(train_df['gleason_score']=='3+4') | (train_df['gleason_score']=='4+3')]['isup_grade'].unique())\n#print(train_df[(train_df['gleason_score']=='3+5') | (train_df['gleason_score']=='5+3')]['isup_grade'].unique())\n#print(train_df[(train_df['gleason_score']=='5+4') | (train_df['gleason_score']=='4+5')]['isup_grade'].unique())","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"#train_df = train_df.drop(columns=['index'])","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Create List for Training"},{"metadata":{},"cell_type":"markdown","source":"The next step is to create two lists for training.  Do not create a list longer than 2750 items long because a list longer than 2750 may result in an error.  In this notebook, a list length of 80 items will be used.  Normally, a list with the length of 2500 to 2750 items was used."},{"metadata":{"trusted":true},"cell_type":"code","source":"train_df = train_df[0:80]","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"image = list(train_df['image_id'])\nlabels = list(train_df['isup_grade'])","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Functions"},{"metadata":{},"cell_type":"markdown","source":"The enhance_image function removes the gray portion from around the prostate biopsy (Zenify)."},{"metadata":{"trusted":true},"cell_type":"code","source":"def enhance_image(slide_path, contrast=1, brightness=15):\n    image = skimage.io.MultiImage(slide_path)[-2]\n    image = np.array(image)\n    img_enhanced = cv2.addWeighted(image, contrast, image, 0, brightness)\n    return img_enhanced","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"The function compute_statistics calculates the portion of white pixels in the region (PAB97)."},{"metadata":{"trusted":true},"cell_type":"code","source":"def compute_statistics(image):\n    \n    width, height = image.shape[0], image.shape[1]\n    num_pixels = width * height\n    \n    num_white_pixels = 0\n    \n    summed_matrix = np.sum(image, axis=-1)\n    # Note: A 3-channel white pixel has RGB (255, 255, 255)\n    num_white_pixels = np.count_nonzero(summed_matrix > 620)\n    ratio_white_pixels = num_white_pixels / num_pixels\n    \n    green_concentration = np.mean(image[1])\n    blue_concentration = np.mean(image[2])\n    \n    return ratio_white_pixels, green_concentration, blue_concentration","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"The functions select_k_best_regions and get_k_best_regions list and select the lowest porportion of white pixels in a particular region (PAB97)."},{"metadata":{"trusted":true},"cell_type":"code","source":"def select_k_best_regions(regions, k=20):\n    regions = [x for x in regions if x[3] > 180 and x[4] > 180]\n    k_best_regions = sorted(regions, key=lambda tup: tup[2])[:k]\n    return k_best_regions","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"def get_k_best_regions(coordinates, image, window_size=512):\n    regions = {}\n    for i, tup in enumerate(coordinates):\n        x, y = tup[0], tup[1]\n        regions[i] = image[x : x+window_size, y : y+window_size, :]\n    \n    return regions","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"The function generate_patches slides over the region to calculate the white pixels then calculates the statistics and then selects the region with the least amount of pixels (PAB97)."},{"metadata":{"trusted":true},"cell_type":"code","source":"def generate_patches(image, window_size=200, stride=128, k=20):\n        \n    max_width, max_height = image.shape[0], image.shape[1]\n    regions_container = []\n    i = 0\n    \n    while window_size + stride*i <= max_height:\n        j = 0\n        \n        while window_size + stride*j <= max_width:            \n            x_top_left_pixel = j * stride\n            y_top_left_pixel = i * stride\n            \n            patch = image[\n                x_top_left_pixel : x_top_left_pixel + window_size,\n                y_top_left_pixel : y_top_left_pixel + window_size,\n                :\n            ]\n            \n            ratio_white_pixels, green_concentration, blue_concentration = compute_statistics(patch)\n            \n            region_tuple = (x_top_left_pixel, y_top_left_pixel, ratio_white_pixels, green_concentration, blue_concentration)\n            regions_container.append(region_tuple)\n            \n            j += 1\n        \n        i += 1\n    \n    k_best_region_coordinates = select_k_best_regions(regions_container, k=k)\n    k_best_regions = get_k_best_regions(k_best_region_coordinates, image, window_size)\n    \n    return image, k_best_region_coordinates, k_best_regions","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"The function glue_to_one_picture glues the 16 patches into one 4X4 image (PAB97)."},{"metadata":{"trusted":true},"cell_type":"code","source":"def glue_to_one_picture(image_patches, window_size=200, k=16):\n    side = int(np.sqrt(k))\n    image = np.zeros((side*window_size, side*window_size, 3), dtype=np.int16)\n        \n    for i, patch in image_patches.items():\n        x = i // side\n        y = i % side\n        image[\n            x * window_size : (x+1) * window_size,\n            y * window_size : (y+1) * window_size,\n            :\n        ] = patch\n    \n    return image","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"WINDOW_SIZE = 128\nSTRIDE = 64\nK = 16","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"The get_i function takes the original image of the biopsy and prepares the image for training."},{"metadata":{"trusted":true},"cell_type":"code","source":"def get_i(image):\n    for i, img in enumerate(image):\n        url = data_dir + img + '.tiff'\n        enhanced_image = enhance_image (url)\n        image, best_coordinates, best_regions = generate_patches(enhanced_image, window_size=WINDOW_SIZE, stride=STRIDE, k=K)\n        glued_image = glue_to_one_picture(best_regions, window_size=WINDOW_SIZE, k=K)\n        glued_image = np.uint8(glued_image)\n        return tensor(glued_image)","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Preparing for Training"},{"metadata":{},"cell_type":"markdown","source":"This part of the code prepares the images for training.  The images are properly labeled.  A batch of images are shown to make sure everything is working properly."},{"metadata":{"trusted":true},"cell_type":"code","source":"blocks = (\n          ImageBlock,\n          CategoryBlock\n          )    \ngetters = [\n           get_i,\n           ColReader('isup_grade')\n          ]\ntrends = DataBlock(blocks=blocks,\n              splitter=RandomSplitter(),\n              getters=getters,\n              item_tfms=Resize(512),\n              )","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"dls = trends.dataloaders(train_df, bs=16)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"dls.show_batch()","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Checkpoints Directory"},{"metadata":{},"cell_type":"markdown","source":"This creates a checkpoint directory for ResNet34 and copies the model to the directory."},{"metadata":{"trusted":true},"cell_type":"code","source":"Path('/root/.cache/torch/checkpoints/').mkdir(exist_ok=True, parents=True)\n!cp '../input/resnet34/resnet34.pth' '/root/.cache/torch/checkpoints/resnet34-333f7ec4.pth'","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Train the Model using ResNet34 and fast.ai"},{"metadata":{"trusted":true},"cell_type":"code","source":"learn = cnn_learner(dls, resnet34, metrics=error_rate)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"torch.cuda.is_available()","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"learn.unfreeze()","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"learn.fit_one_cycle(3)","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Predict from Train Dataset"},{"metadata":{"trusted":true},"cell_type":"code","source":"interp = ClassificationInterpretation.from_learner(learn)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"predictionIndex = []\nnum = len(learn.dls.train_ds)\n\nfor i in range(num):\n  predictionIndex.append (learn.predict(learn.dls.train_ds[i][0]))","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"singlepredictionIndex = []\nnum = len(predictionIndex)\n\nfor i in range(num):\n  singlepredictionIndex.append (predictionIndex[i][0])","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"prediction_df = learn.dls.train_ds.items","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"prediction_df = prediction_df.assign(Prediction=singlepredictionIndex)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"print (prediction_df)","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Remove High Error Samples"},{"metadata":{},"cell_type":"markdown","source":"Remove biopsies with a high error difference. The error difference will be at least two ISUP grades higher than the assigned ISUP."},{"metadata":{"trusted":true},"cell_type":"code","source":"indexnames35 = prediction_df[(prediction_df['isup_grade'] == 3) & (prediction_df['Prediction'] == '5')].index\nindexnames05 = prediction_df[(prediction_df['isup_grade'] == 0) & (prediction_df['Prediction'] == '5')].index\nindexnames15 = prediction_df[(prediction_df['isup_grade'] == 1) & (prediction_df['Prediction'] == '5')].index\nindexnames13 = prediction_df[(prediction_df['isup_grade'] == 1) & (prediction_df['Prediction'] == '3')].index\nindexnames20 = prediction_df[(prediction_df['isup_grade'] == 2) & (prediction_df['Prediction'] == '0')].index\nindexnames25 = prediction_df[(prediction_df['isup_grade'] == 2) & (prediction_df['Prediction'] == '5')].index\nindexnames40 = prediction_df[(prediction_df['isup_grade'] == 4) & (prediction_df['Prediction'] == '0')].index\nindexnames31 = prediction_df[(prediction_df['isup_grade'] == 3) & (prediction_df['Prediction'] == '1')].index\nindexnames41 = prediction_df[(prediction_df['isup_grade'] == 4) & (prediction_df['Prediction'] == '1')].index\nindexnames35 = prediction_df[(prediction_df['isup_grade'] == 3) & (prediction_df['Prediction'] == '5')].index\nindexnames05 = prediction_df[(prediction_df['isup_grade'] == 0) & (prediction_df['Prediction'] == '5')].index\nindexnames15 = prediction_df[(prediction_df['isup_grade'] == 1) & (prediction_df['Prediction'] == '5')].index\nindexnames13 = prediction_df[(prediction_df['isup_grade'] == 1) & (prediction_df['Prediction'] == '3')].index\nindexnames20 = prediction_df[(prediction_df['isup_grade'] == 2) & (prediction_df['Prediction'] == '0')].index\nindexnames25 = prediction_df[(prediction_df['isup_grade'] == 2) & (prediction_df['Prediction'] == '5')].index\nindexnames40 = prediction_df[(prediction_df['isup_grade'] == 4) & (prediction_df['Prediction'] == '0')].index\nindexnames31 = prediction_df[(prediction_df['isup_grade'] == 3) & (prediction_df['Prediction'] == '1')].index\nindexnames41 = prediction_df[(prediction_df['isup_grade'] == 4) & (prediction_df['Prediction'] == '1')].index\nindexnames30 = prediction_df[(prediction_df['isup_grade'] == 3) & (prediction_df['Prediction'] == '0')].index\nindexnames02 = prediction_df[(prediction_df['isup_grade'] == 0) & (prediction_df['Prediction'] == '2')].index\nindexnames14 = prediction_df[(prediction_df['isup_grade'] == 1) & (prediction_df['Prediction'] == '4')].index\nindexnames50 = prediction_df[(prediction_df['isup_grade'] == 5) & (prediction_df['Prediction'] == '0')].index\nindexnames03 = prediction_df[(prediction_df['isup_grade'] == 0) & (prediction_df['Prediction'] == '3')].index\nindexnames04 = prediction_df[(prediction_df['isup_grade'] == 0) & (prediction_df['Prediction'] == '4')].index\nindexnames51 = prediction_df[(prediction_df['isup_grade'] == 5) & (prediction_df['Prediction'] == '1')].index\nindexnames53 = prediction_df[(prediction_df['isup_grade'] == 5) & (prediction_df['Prediction'] == '3')].index\nindexnames24 = prediction_df[(prediction_df['isup_grade'] == 2) & (prediction_df['Prediction'] == '4')].index\nindexnames42 = prediction_df[(prediction_df['isup_grade'] == 4) & (prediction_df['Prediction'] == '2')].index\nindexnames52 = prediction_df[(prediction_df['isup_grade'] == 5) & (prediction_df['Prediction'] == '2')].index","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"prediction_df.drop(indexnames35, inplace=True)\nprediction_df.drop(indexnames05, inplace=True)\nprediction_df.drop(indexnames15, inplace=True)\nprediction_df.drop(indexnames13, inplace=True)\nprediction_df.drop(indexnames20, inplace=True)\nprediction_df.drop(indexnames25, inplace=True)\nprediction_df.drop(indexnames40, inplace=True)\nprediction_df.drop(indexnames31, inplace=True)\nprediction_df.drop(indexnames41, inplace=True)\nprediction_df.drop(indexnames30, inplace=True)\nprediction_df.drop(indexnames02, inplace=True)\nprediction_df.drop(indexnames14, inplace=True)\nprediction_df.drop(indexnames50, inplace=True)\nprediction_df.drop(indexnames03, inplace=True)\nprediction_df.drop(indexnames04, inplace=True)\nprediction_df.drop(indexnames51, inplace=True)\nprediction_df.drop(indexnames53, inplace=True)\nprediction_df.drop(indexnames24, inplace=True)\nprediction_df.drop(indexnames42, inplace=True)\nprediction_df.drop(indexnames52, inplace=True)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"print (prediction_df)","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Export Dataframe"},{"metadata":{},"cell_type":"markdown","source":"For the first dataframe, the .csv file was called prediction_df.csv.  For the second dataframe, the .csv file was called train_df.csv."},{"metadata":{"trusted":true},"cell_type":"code","source":"prediction_df.to_csv('example_df.csv')","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Works Cited"},{"metadata":{},"cell_type":"markdown","source":"PAB97. “Better image tiles - Removing white spaces.” Kaggle, 22 May 2020, www.kaggle.com/rftexas/better-image-tiles-removing-white-spaces.\n\n“Prostate CANcer GraDe Assessment (PANDA) Challenge.” Kaggle, www.kaggle.com/c/prostate-cancer-grade-assessment/overview/description. \n\nrmicrobe. “Microanatomy of the Prostate.” Kaggle, 11 June 2020, www.kaggle.com/rmicrobe/microanatomy-of-the-prostate. \n\nZac Dannelly. \"PANDA_Analysis.\" Kaggle, 14 May 2020, www.kaggle.com/dannellyz/panda-analysis.\n\nZenify. “Let's Enhance the Images!” Kaggle, 03 May 2020, www.kaggle.com/debanga/let-s-enhance-the-images."}],"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}