{"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"},"kaggle":{"accelerator":"none","dataSources":[{"sourceId":97984,"databundleVersionId":14096757,"sourceType":"competition"}],"isInternetEnabled":true,"language":"python","sourceType":"notebook","isGpuEnabled":false}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"using filtering or advanced image\nprocessing techniques to separate the signal from the back ground grind\ncorrect for skewing:-\naccount rotation&mmisalignment introduced \nduring scanning or photography.\nhandleimage artifacts:- \nadrees real-world issue like wrinkles,sniats\n& shadows that appear on the images \nimage augmentation :- systematically generated additional training data by\napplying transformation like scaling,rrotation,sgifts to make \nthis model more to bust\nmodel architecture:-\n•Encoder-decoder set up :-\nCommon approach is using an encoder-decoder architecture like au-net\n•Encoder:-A convolutional neutral network (CNN)\nprocess the processed image & learned \nto extract important features essentially \nunderstanding the context of the ECG graph \n•Decoder:- the decoder then takes these features & reconstructs the time series signal \nthis part of our code uses the out put from our model to create\nthe final time series data\n•Extract signal from mask:-\nconvert the signal segmentation mask \nproduced by our model into a series  of data points\nthis could involve finding the vertical position of the signalfor each horizontal pixel column\n•calibrate using image grid :-\nare the ECG grid information removed during preprocessing \nto properly scale the amplitude \n•Segment leads:-\nExtract the individual 12 leads the \nful image (or) organize them into the correct order\n•Noisefiltering :- smooth out any remaining\nnoise or artifacts in the digitized\nsignal using a signal proceeding filter\n•Aligin with ground thruth:-\nmatch the predicted signal with the ground\ntruth time series to account for minor shifts ,as is required by competition evaluation metric\n●thanks for giving me\n    this\nopurtunities\n","metadata":{"trusted":true},"outputs":[],"execution_count":null}]}