Upgraded user service

The main feature in this release is a new authentication service. This service combines basic auth (for API usage) with session cookie-based auth (for web clients). Session cookies is (marginally) safer than basic auth since username and passwords is not communicated with each request. Instead, a session cookie is obtained by logging in. The session cookie is tied to a user, user agent and IP address until the user logs out or the session expires. If the session cookie (containing a random string) is sent with a request, the user is authenticated.

Other features in this release is

• smoother sending of images, the sorting order from the list is preserved, possible to preserve anonymous patient names etc
• added bulk operations for listing all images for a patient or study, and for deleting a list of images
• added advanced filtering to queries
• possibility to upload a jpeg image which is wrapped in a dicom object. Including viewing of such images

Compression and Encryption

In this release we have added the possibility to encrypt all traffic using SSL/TLS. The README file describes the steps necessary to install SSL.

We have also added compression of datasets before transfer to increase transfer rates and lower network load.

DICOM hierarchy simplifications

This release contains small changes with a big impact. We removed the ImageFile entity (connected to each DICOM Image entity) and no longer persist the file name and source of each image. The filename is now derived directly from each Image.

Further, we changed how uniqueness is defined for Study:s, Series and Images. Before, these were defined using StudyInstanceUID, SeriesInstanceUID and SOPInstanceUID respectively. This could (and sometimes did) lead to inconsistencies in the DICOM hierarchy when adding a new file, for instance when a new file lead to adding a new Patient to the system, but the Study was already added from before and connected to another Patient. We now define uniqueness as the xxInstanceUID properties paired with the parent id. For instance, this means that a Series is defined by its SeriesInstanceUID paired with its study id.

Forwarding service

This release adds a forwarding service in which a source (SCP, directory, box, user or similar) can be connected to a destination (currently either SCU or box). For each image added to slicebox, the forwarding service will match the source of the image with the set of specified forwarding rules. If there is a match, slicebox will automatically forward the image to the destination that the forwarding rule specifies. There is also an option whether to keep the images after forwarding, or delete them.

Series types, tags and advanced filtering

This release adds an integration interface with applications using the concept of series types - labels attached to series that signifies the kind of examination or procedure they are part of. The assignment of series types is based on rules made up of combinations of values for DICOM attributes. For instance, a series with Modality=CT, and BodyPartExamined=HEAD can be assigned to the series type "Head examination (CT)". The assignment of series types is automatic and occurs in the background as soon as series type rules are updated. When browsing for images, results can be filtered on one or more series types.

In addition to series types, there are series tags. Users can add and delete series tags as they see fit to series to group them.

Filtering works with any combination of series types, series tags and series sources.

Import, export

This release adds convenient functions for importing image data through the slicebox client by dragging-and-dropping files and folders into an upload are in the new import view, and exporting image data as zip archives using a new object action available on patients, studies and series.

This release also includes a bug fix that made image transfers in one direction very slow.

Anonymization Service

This release includes some refactoring on the server side. There is now an anonymization service in charge of anonymization as well as harmonizing anonymization such that a certain patient is consistently anonymized between separate sends or manual anonymizations. Further, the service takes care of reversing anonymization for incoming images belonging to patients which originates from that instance. This makes it possible for a collaborator to return a new series (e.g. a secondary capture) which will be added to the correct patient on receive.

There is now an option to force anonymization of images on your instance, useful if you happened to import images with patient information that you wish to get rid of.

Box to box communication with previous releases of slicebox is not possible. Make sure all instances run this version or newer, until further notice.

Accession Number fix

Random accession numbers in anonymized files are now of length 16 characters to fit withing the VR type SH.

Anonymization bug fix

Fixed a bug where anonymization added new tags in for some datasets.

Sources and source filtering

Slicebox now keeps track of the origin of each dataset, be it from a watched directory, a PACS system, a remote box or through an API call by a logged in user. Meta data can now be filtered by source type and source id, making it possible to restrict users to data from a particular source.