![](\"https://content.presspage.com/uploads/2110/1920_gettyimages-1216304354.jpg\")
\n\n![](\"https://siim.org/resource/resmgr/SIIM_logo-600x315.png\"){kind=logo}
\nSociety for Imaging Informatics in Medicine (SIIM) is the leading healthcare professional organization for those interested in the current and future use of informatics in medical imaging. The society's mission is to advance medical imaging informatics across the enterprise through education, research, and innovation in a multi-disciplinary community.
\n\nDigital Imaging and Communications in Medicine (DICOM) is the standard for the communication and management of medical imaging information and related data. DICOM is most commonly used for storing and transmitting medical images enabling the integration of medical imaging devices such as scanners, servers, workstations, printers, network hardware.
\n\n Read more about DICOM here\n\nFollowing are some of applications of using DICOM format in medical imaging
\n\n- Collaboration With Existing IT Systems\n- High-Performance Review\n- Complete Scanning and Image Reviewing\n\n Read more about DICOM applications here","metadata":{}},{"cell_type":"code","source":"##---------------------------------------\n#dicom to pixel array converting function\n##---------------------------------------\n\n# Ref : https://www.kaggle.com/raddar/convert-dicom-to-np-array-the-correct-way\ndef dicom2arr(path, voi_lut = True, fix_monochrome = True):\n dicom = pydicom.read_file(path)\n \n if voi_lut:\n arr = apply_voi_lut(dicom.pixel_array, dicom)\n else:\n arr = dicom.pixel_array\n \n if fix_monochrome and dicom.PhotometricInterpretation == \"MONOCHROME1\":\n arr = np.amax(arr) - arr\n \n arr = arr - np.min(arr)\n arr = arr / np.max(arr)\n arr = (arr * 255).astype(np.uint8)\n \n return arr\n\n##-----------------------\n#resizing the pixel array\n##-----------------------\n\ndef resizeArr(arr,is_train=True):\n im = Image.fromarray(arr)\n if is_train:\n train_w.append(im.size[0])\n train_h.append(im.size[1])\n else:\n test_w.append(im.size[0])\n test_h.append(im.size[1]) \n im = im.resize((HEIGHT,WIDTH),resample= Image.LANCZOS)\n return im\n\n##------------------------------\n#Filename for resized jpg images \n##------------------------------\n\ndef getFilename(filepath,is_train=True):\n '''\n Fromat = '{STUDY-ID}_{SUB-STUDY-ID}_{IMAGE-ID}.jpg'\n '''\n ls = filepath.split(\"/\")\n filename = ls[-3]+'_'+ls[-2]+'_'+ls[-1].split(\".\")[0]+\".jpg\"\n \n if is_train:\n train_id.append(ls[-1].split(\".\")[0])\n else:\n test_id.append(ls[-1].split(\".\")[0])\n return filename\n\n##-------------------\n#Finally saving image\n##-------------------\n\ndef saveImage(filepath,mainpath=FINAL_TRAIN_PATH,train=True):\n arr = dicom2arr(filepath)\n arr = resizeArr(arr,train)\n path = os.path.join(mainpath,getFilename(filepath,train))\n \n arr.save(path)","metadata":{"execution":{"iopub.status.busy":"2021-05-23T02:08:53.0346Z","iopub.execute_input":"2021-05-23T02:08:53.03485Z","iopub.status.idle":"2021-05-23T02:08:53.04679Z","shell.execute_reply.started":"2021-05-23T02:08:53.034826Z","shell.execute_reply":"2021-05-23T02:08:53.045611Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"os.makedirs(FINAL_TRAIN_PATH,exist_ok=True)\nos.makedirs(FINAL_TEST_PATH,exist_ok=True)\n\n#Preparing Training Image Data\nfor filepath in TRAIN_FILES:\n saveImage(filepath)\n\n#Preparing Test Image Data\nfor filepath in TEST_FILES:\n saveImage(filepath,mainpath=FINAL_TEST_PATH,train=False)\n \n# Ref : https://www.kaggle.com/xhlulu/siim-covid-19-convert-to-jpg-256px\n!tar -zcf train.tar.gz -C \"./dataset/train/\" .\n!tar -zcf test.tar.gz -C \"./dataset/test/\" .","metadata":{"execution":{"iopub.status.busy":"2021-05-23T02:08:53.048363Z","iopub.execute_input":"2021-05-23T02:08:53.049045Z","iopub.status.idle":"2021-05-23T02:09:00.338909Z","shell.execute_reply.started":"2021-05-23T02:08:53.048982Z","shell.execute_reply":"2021-05-23T02:09:00.33776Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"![](\"https://libcst.readthedocs.io/en/latest/_images/graphviz-d27e3495fa9bb130d76879db599060e8039a9fc5.png\")
","metadata":{}},{"cell_type":"code","source":"train_df = pd.merge(image_df,study_df,on = \"StudyInstanceUID\") # Merging study_df and image_df\n\ntry :\n train_df = pd.merge(train_df,meta_train,on = \"ImageInstanceUID\") # Merging to meta_train for height,width\nexcept:\n pass\n\n# Filling NaN values \ntrain_df[\"boxes\"].fillna(\"[{'x':0,'y':0,'width':1,'height':1}]\",inplace=True)\ntemp = train_df # for going through the data\ntrain_df[\"boxes\"] = train_df[\"boxes\"].map(lambda x : ast.literal_eval(x))\n\ntry:\n columns = [\"ImageInstanceUID\",\"StudyInstanceUID\",\"label_id\",\"study_label\",\"height\",\"width\",\"boxes\",\"image_label\"] # for proper order\n train_df = train_df[columns]\nexcept:\n columns = [\"ImageInstanceUID\",\"StudyInstanceUID\",\"label_id\",\"study_label\",\"boxes\",\"image_label\"] # for proper order\n train_df = train_df[columns]\n\ntrain_df.to_csv(\"./train.csv\",index=False)\ntrain_df.head()","metadata":{"execution":{"iopub.status.busy":"2021-05-22T17:05:23.866546Z","iopub.execute_input":"2021-05-22T17:05:23.866909Z","iopub.status.idle":"2021-05-22T17:05:24.12265Z","shell.execute_reply.started":"2021-05-22T17:05:23.866878Z","shell.execute_reply":"2021-05-22T17:05:24.121203Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import shutil\nshutil.rmtree(\"./dataset\",ignore_errors=False)","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"![](\"https://external-preview.redd.it/vWYcdynWxuFy6bKYpZwuw6KiNgYuBPM6daHCwWRs4mo.png?auto=webp&s=fe3bf857b3c9f1369aed1ea0bc5e1acd8ae39449\")
","metadata":{}}]}