Fork me on GitHub

Browser Area

What is it for?

The browser area allows you to discover and import data into medInria. medInria maintains a database of the available data, which you can populate from external files, or other sources added by plugins. For instance, a Dicom plugin, or a Shanoir plugin are included in the medInria 2.0.0 release.

Source Selector

The Source Selector offers a choice of tabs listing the data available in different sources. The sources present by default are :

  • Database: The database of data indexed or imported into medInria. It references all data available directly to the viewer area. This may include temporary data. Temporary data are data generated by some algorithm, or temporary imported just for the duration of one session of medInria. The data can be removed by right clicking and selecting Remove.
  • File System: This lists the data available on the local filesystem. Here new data can be imported indexed, or selected for viewing by right-clicking.

The other sources added by additional plugins, present in the release are for the moment:

  • Dicom: Allows you to read Dicom Dirs on the local file system (like cds), or to interrogate a PACS.
  • Shanoir: Queries a Shanoir ( database.

File System

What is import, view, load and index ?

There are several ways to access the data, depending on how you wish to use it. These can be accessed by right-clicking on a file or directory in the FS representation.

Import - This causes medInria to take a copy of the selected data into its local database. This allows you to view the data faster (since medInria optimises the storage for fast loading), but changes to the original file will not be reflected in medInria's copy. The data will be available the next time you open medInria, even if the original file has been removed.

Load - The selected data is loaded into the temporary database. The data is not copied, and will not be available in the database after medInria is closed. Temporary data is shown in italics in the DB.

View - The selected data is loaded into temporary storage, as with the previous option, but also a new Visualisation configuration is opened showing the data. The data is not imported, and will not be available in the database when medInria is closed. 

Index - An entry for the selected data is created in medInria's permanent database. Contrary to the import option, the actual data is not copied, so the data must be available when you use it. There is also no optimisation of storage: ie volumes from Dicom slices will not be stored, the reconstruction will have to be done every time you reload this data from the database.

Which formats are readable?

The answer to this question depends on the readers and writers in your plugins. By default, the included readers are all the formats readable by itk (,  most DICOM images, vtk meshes, adn composite dataset made of sevral heterogenous files (see composite DTI data)

How to save temporary loaded data to the database?

(or how to save the image returned by a process)

After loading images in the database (e.g. either by right clicking "Load" in the File Browser or by creating new images resulting of image processes) there are 3 different ways to import them into the database :

  •  in the Browser workspace : by right clicking "Save" the selected items
  •  in Diffusion/Filtering/Registration workspaces : by clicking any button like "Store in Database"
  •  when attempting to close the application (by pressing the "Quit" button) while non-persistent data have been created : a dialog menu proposes you to save unsaved data to the medInria database before exiting.

How to bookmark a location?

medInria allows bookmarking of folders, so as it is easier to access it in the future. In order to do so, the user must right-click on the desired folder and select "Bookmark". Right after, the new location should appear just below the "Bookmarks" section in the Fs tab of the Source Selector, as shown in the picture:


What is the database viewer? 

Once imported, indexed, or loaded, data is shown in the Database (Db) viewer, which is a tree structure organised by a Patient/Study/Series hierarchy, following the DICOM precepts. In italics are the temporary data. Interactions with the B viewer are:

  • Double clicking on a row switches to the Visualization Configuration in the Viewer area.
  • right clicking on a row allows you to perform these actions on a row:
    • view
    • export
    • delete

If a row is a Study, the action will be applied to all the series below it, and if it is a Patient, to all its studies, and as result series. 

How to remove?

Removing an item from the database is as easy as importing or indexing it. Right-click on the item you want to remove in the database view (either just a series or a whole patient) and select "Remove" from the context menu, as shown in the following picture.

medInria will ask you for confirmation, as after an image is removed, it cannot be undone.
Note that the removal of many images, or a patient with many studies/series, might take a while.



How to export?

After right clicking on an entry, the export option will open a file dialog. The extension you give to the file name will determine the writer used to produce the exported data. If there is no writer capable of writing this extension, the export will fail.

How to Filter?

Filters can be used to reduce the size of the displayed data. Filters are added by selecting a field name. Only one field of each type is possible, you can delete a filter by selecting its name in the drop down list of fields and clicking on the minus sign.


The tree can also be sorted (in [anti]lexicographic order) by clicking on the column headers.


DICOM data source

The DICOM datasource looks at a DICOM dir as if it was a small database, and allows queries on remote PACS systems. The top row gives you search criteria: Patient name, Study descritption, Series description, modality, gender, and filters between two dates.

Patients, studies and series will be displayed in 3 different lists, Studies and Series being updated given the selected patient and study. 

SCU configuration

medInria acts as a SCU: Service Class User in the DICOM terminology. Prior to any DICOM query to a SCP (Service Class Provider) the application needs to identify itself as a SCU. This is done in the top right tool box. The user must specify :

  • AE Title, that will identify the SCU on the network
  • The TCP incoming port. The port that will be used by the SCP to send data.
  • The hostname, that will allow the SCP to locate the machine. As any TCP-IP hostname it can be an IP address or a name known to a DNS server.

Server configuration

The tool box below, called DICOM servers settings, holds a list of registered known SCPs. Before saving an entity, the user must fill in server hostname, AE Title, TCP port and the name it will be listed by in the top left list of servers.

There is an echo button in order to test a value before saving it. It sends the DICOM echo message to DICOM servers.

Importing from DICOM services to the local DB

The next two tool boxes underneath will give some information useful before importing any data. The last tool box brought by the dicom plugin is the import in database one. It imports the selected series into the database. Notice that it is not possible to index or load the data, only import it at the moment.



SHANOIR data source

The SHANOIR data source will query the database on a SHANOIR server. The topmost tool box will do all the setup with a username/password pair for authentication, and a server/Port pair for the remote server.

The Queries tool box has search fields such as study name, patient name, date and dataset name. 

As for the DICOM data source, only full importation is possible.



Importing composite DTI data ? 

WARNING: Currently composite data sets is just a technological preview. The datasets can be imported and saved in the database but cannot be used directly from any other medInria plugin from the time being. It is intended to add functionality to the next medInria release to create and use these composite datasets from some plugins.

A composite data set is a way to associate multiple data together. Take for example, a raw DTI data set. Such a dataset contains several images and each of these images is associated to a gradient direction (which is not always available in the image meta-data). As a user, it would be nice to manipulate all these images and their associated gradient as a whole. This is true for the visualization plugins, the processing plugins and also for the database (you usually do not want to have one entry for each of the tens of images which constitute the dataset).

Technically, a composite dataset is simply a zip container that gathers all the needed data in a single file. Typically, the zip will be expanded to a directory containing the individual files and a simple text file that will list them and associated the needed metadata.

Here is for example, the structure of a composite dataset made for DTI images:

unzip -l dwi.cds

Length      Date    Time    Name
---------  ---------- -----   ----
        0  08-14-2011 02:01   dwi/
  1900842  06-21-2011 14:36   dwi/wholedataset-0.mha
  1900842  06-21-2011 14:36   dwi/wholedataset-6.mha
  1900842  06-21-2011 14:36   dwi/wholedataset-5.mha
  1900842  06-21-2011 14:36   dwi/wholedataset-2.mha
  1900842  06-21-2011 14:36   dwi/wholedataset-3.mha
  1900842  06-21-2011 14:36   dwi/wholedataset-4.mha
      272  08-14-2011 02:01   dwi/Description.txt
  1900842  06-21-2011 14:36   dwi/wholedataset-1.mha
---------                     -------
 13306166                     9 files

This simple example shows two things, the dwi.cds is a simple zip file. Consequently, you can continue to manipulate the data it contains with other tools (just unzip it, change the files and rezip it). Second, the file dwi.cds is simply expanded into a directory dwi that contains the various images plus a metadata text file named dwi/Description.txt. Here is the content of this file:

# A DTI dataset

Images: 7 MetaData: 0
"wholedataset-0.mha" [0 0 0]
"wholedataset-1.mha" [1 0 1]
"wholedataset-2.mha" [-1 0 1]
"wholedataset-3.mha" [0 1 1]
"wholedataset-4.mha" [0 1 -1]
"wholedataset-5.mha" [1 1 0]
"wholedataset-6.mha" [-1 1 0]

In this example, we see that the Description.txt file is a simple text file. Blank lines are ignored, Comments start with the '#' character and run till the end of the line. The first line is special and acts as a magic cookie to identify a composite dataset (the "# MEDINRIA COMPOSITE DATA:") followed by the type of composite data (here "DWI" for diffusion weighted images) and a version (currently unused). Everything else is specific to the DWI file format, here it is just stated that there are 7 images and special metadata section (metadata are directly associated to the images). Then, there is a list of image names (contained in the dwi directory) and their associated gradients.

The whole dwi.cds containing this information can be imported as a whole in the database (currently this is the only operation permitted) as any other file. See section "How to import".


Back to Documentation Index