qxiaobu / flannel Goto Github PK

While reviewing the metadata.csv file for the covid-chestxray-dataset-master information, it appears that are many classifications that are not accounted for. While some make sense, others would appear that they should be included. This could be due to a change in the format of this file; however, could you confirm this is intended?

In the image below, the distinct findings are shown on the right hand side. The yellow highlighted items are forms of pneumonia which would not get picked up by the python script's logic. As they are not included in the lists defined in the processing script.

This code creates a series of lists whose elements are later compared to the finding column from the metadata.csv file.

Unfortunately the logic does not work, which I noticed when it reported 0 COVID images.

The issue is that the information from the metadata.csv file is not a single item which neatly aligns with a member of the list, rather it is a hierarchy from least to most specific. For example, COVID-19 isn't referenced as COVID-19, it is: 'Pneumonia/Viral/COVID-19'. Because of this, the logic fails.

The "easy" fix is to just split the string by / and use the right-most entry.

After this update, x0 - x4 appear to give meaningful results

Hi,
It appears that in the ensemble_step1.py code, some code is missing. For eg: it does not have the training steps. Is that a correct observation.
Thanks,

Looking to clarify if a filter restricting process only to Views of 'AP' is intentional or an oversight.

This can be found in the main processing loop:

I ask if this was intentional due to the fact that later processing code looks for 'PA' views. Due to the filter above, there will never be any PA items.

Could you please confirm if this is intended functionality?

The column numbers referenced in the get_covid_data_dict.py routine do not seem to align with the metadata.csv file for the covid-chestxray-dataset-master. Perhaps the xray data has changed? Could use some pointers as to what subject_id & disease correspond to.

Is
subject_id = offset
disease = finding

?