Import

Importing datasets into your current project can be done by clicking on the + button in the Database tab, which opens the Import dialog. The dialog has four tabs on the top left:

The DICOM tab is used for importing individual dcm files or DICOM folders. In this tab you can specify a folder (local or on the network) to be scanned. As you begin the scan, the table below will fill up with Studies, Patient and Series information. Various DICOM properties can be inspected in the panel on the right.

When pressing Import, all the Series that have a checkmark will be imported into the database.
The Other Files tab is used for importing regular formats such as JPG, PNG, NIFTI, MHD or ImFusion files (IMF). As you navigate the folders, the files that can be loaded will be shown in the middle. Marking and unmarking files for import is done by simply clicking on them. Files that are incompatible with the project are shown in red and cannot be imported.

If Recursive is checked, all the directories below the current one will be traversed to find datasets candidates for import.

The Hide incompatible checkbox lets you hide incompatible (shown in red) datasets.

The Check All and Uncheck All respectively lets you mark and unmark all the currently shown datasets for import.
The Project tab is used for importing data from another project into the current. Depending on which options are selected, this can be used to add new datasets or update the existing ones (if they’ve been imported in such a way before). If the types of the two project match, it can optionally be used to update the annotations of existing datasets.
The Data List tab is used to import a set of data (and optionally associated labels) as a single operation. The expected format for the data list is a text file with the full path to a dataset on each line.

If a label is already available for a dataset, you can indicate its path on the same line, separated with a semi-colon (;). For segmentation projects, a valid label is an image encoded as unsigned bytes with the same dimension as the image.

For instance, here is a sample data_list.txt that can be used:
```
C:/Data/MyDataset/Image1.mhd;C:/Data/MyDataset/Labels1.mhd
C:/Data/MyDataset/Image2.mhd;C:/Data/MyDataset/Labels2.mhd
C:/Data/MyOtherDataset/Images/0.nii;C:/Data/MyOtherDataset/Labels/2.nii
C:/Data/MyOtherDataset/Case1.nii
C:/Data/MyOtherDataset/Case2.nii
```

In all tabs, the dropdown next to the Import button lets you select how you want the data to be imported:

If Link to original location is selected, the project will only record the path of the data that you imported. Anytime you try to open the dataset, it will open the files from their original location.
If Copy datasets to project is selected, the datasets are copied into the project during import. There are several pros and cons to this, such as:
- Copied data is self-contained and therefore easily portable
- Copied datasets may load faster (for example, if the original data is on a remote server)
- The original data can be moved, renamed… without issues
- The project uses more disk space
- Importing takes more time

After importing data, a small report will tell you how many datasets were successfully imported, and whether some of the datasets were skipped and why.

Note

It is also possible to directly import data by dragging and dropping a single file or folder into the Database tab.

Note

The folder scanning and import operation can take some time, because it needs to explore all files contained in the directory and try to load them. However, this is a one-time operation: once the images are part of the project, any operation on them will be much faster.

Importing Segmentations

DICOMs with a modality of SEG are not imported as regular datasets but as segmentations for a particular dataset. A standard conforming DICOM Segmentation file contains a list of unique identifiers (UIDs) that reference the dataset the segmentation belongs to. ImFusion Labels will use those UIDs to find the correct dataset in the database. This only works with datasets that were imported from DICOM. Whether the datasets have been copied to the project or still link to the original DICOMs is not relevant.

If your project has multiple layers, you will see an additional dropdown next to the Import button. It lets you choose whether all layers stored in the DICOM should be imported or only the currently active layer.

The labels are stored by name in the DICOM Segmentation file. In case your project uses duplicate names for labels, only the first label with a matching name will be imported. Since this might lead to unexpected results, duplicated label names should be avoided in general.