Setup¶

Complete Bob’s installation instructions. Then, to install this package, do this:

$ conda activate <myenv>
(<myenv>) $ conda install bob.med.tb

Note

The value <myenv> should correspond to the name of the environment where you initially installed your Bob packages.

Datasets¶

The package supports multiple chest X-ray datasets, but does not include the raw data itself, which you must procure.

To setup a dataset, do the following:

Download the dataset from the authors website (see Supported Datasets for download links and details), unpack it and store the directory leading to the uncompressed directory structure.

Warning

Our dataset connectors expect you provide “root” paths of raw datasets as you unpack them in their pristine state. Changing the location of files within a dataset distribution will likely cause execution errors.

For each dataset that you are planning to use, set the datadir to the root path where it is stored. E.g.:

(<myenv>) $ bob config set bob.med.tb.montgomery.datadir "/path/to/montgomery"

To check supported raw datasets and your current setup, do the following:

(<myenv>) $ bob tb dataset list
Supported datasets:
- montgomery: bob.med.tb.montgomery.datadir = "/Users/yourname/work/bob/dbs/montgomery"
* montgomery_RS: bob.med.tb.montgomery_RS.datadir (not set)
- shenzhen: bob.med.tb.shenzhen.datadir = "/Users/yourname/work/bob/dbs/shenzhen"
* shenzhen_RS: bob.med.tb.shenzhen_RS.datadir (not set)
- indian: bob.med.tb.indian.datadir = "/Users/yourname/work/bob/dbs/indian"
* indian_RS: bob.med.tb.indian_RS.datadir (not set)
- nih_cxr14_re: bob.med.tb.nih_cxr14_re.datadir = "/Users/yourname/work/bob/dbs/NIH_CXR14"

This command will show the set location for each configured dataset, and the variable names for each supported dataset which has not yet been setup.

Note

Tuberculosis datasets having radiological signs in place of images are recognizable by the “_RS” at the end of their name and do not need a location to be set.

To check whether the downloaded version is consistent with the structure that is expected by this package, run bob tb dataset check <dataset>, where <dataset> should be replaced by the dataset programmatic name. E.g., to check Montgomery files, use:
```
(<myenv>) $ bob tb dataset check montgomery
...
```
If there are problems on the current file organisation, this procedure should detect and highlight which files are missing (cannot be loaded).