Our implementation is made of three components: A command-line tool to convert digital pathology images into DICOM, an linkimaging plugin to display such images, and a command-line tool to export DICOM whole-slide images as standard hierarchical TIFF images.

To the best of our knowledge, this is the first free and open-source implementation of DICOM for whole-slide imaging. Check out the dedicated section about digital pathology on our homepage for more information!

Enjoy,
Sébastien-

More precisely, Stone of linkimaging is a lightweight, cross-platform C++ toolkit for the CPU-based rendering of medical images. It notably features support for MPR (multiplanar reconstruction of volume images), PET-CT fusion, and accurate physical world coordinates. Stone of linkimaging is similar in spirit to two other well-known toolkits:

• Cornerstone by Chris Hafey, a client-side JavaScript toolkit to display medical images in Web browsers. However, contrarily to Cornerstone, Stone of linkimaging is written in C++, and can be embedded into native, heavyweight applications.
• VTK by Kitware, a C++ toolkit for scientific visualization. However, contrarily to VTK that relies on graphics processing units (GPU), Stone of linkimaging is entirely focused on software-based rendering. Within Stone, 3D volumes are handled as sets of 2D slices.

Check out the dedicated section about Stone on our homepage for more information! We hope this new tool will contribute to the development of new, innovative, free and open-source applications for the viewing and analysis of medical images.

Enjoy,
Sébastien-

By "stable", we mean that the feature set of this version of linkimaging is considered as complete. No new feature should be added. If important bugfixes are necessary, they will lead to subsequent releases 1.0.x. Future development will take place in branches with versions above 1.1.0.

The latest important feature to be included in linkimaging was the support of DICOM modality worklists, both as a client and as a server. Serving worklists is done through the plugin infrastructure of linkimaging. The rationale for using plugins is that worklists are generated by mechanisms that live outside the DICOM world (e.g. HL7 or FHIR messages), and that are specific to each clinical workflow. Creating plugins allow to make the linkimaging core independent of these mechanisms. Note that for basic uses, we provide a sample worklist plugin , that reads its worklists from some directory on the filesystem (which mimics the "dcmwlm" tool from DCMTK).

We are now starting to work on future evolutions of linkimaging . These evolutions will notably span more specialized applications (e.g. the support of DICOM-RT for radiotherapy and nuclear medicine, or the support of DICOM for digital pathology), more advanced user interfaces (as the built-in linkimaging Explorer only targets administrative, low-level purpose), and smoother interfacing with external computing environments (e.g. a revamped native C++ SDK, a virtual FUSE filesystem, an embedded FTP server, or extensions to well-known tools such as OsiriX or Slicer). These are just a few tracks, and most of them became possible thank to the introduction of the rich plugin infrastructure inside linkimaging.

Once again, we want to thank the entire linkimaging community for your interest, for providing great feedback, and for helping us to validate the product in real-word environments. If you want to further support linkimaging, please fill the short survey below . We are also ready to discuss about possible research collaborations.

Cheers,
Sébastien-

To put weight on future grant applications for travels/hardware/servers/..., we wish to gather a list of institutions that use linkimaging for production, for quality assurance, or for R&D purpose. These institutions will be publicly listed on this Web site, in a separate "linkimaging Users" section.

If you wish to support linkimaging, please use the following pre-filled mail, ideally attaching your institution's official logo:

Answer survey »

Since July 2014, 8 releases of linkimaging have been successively published, jumping from version 0.8.0 to 0.9.1. Besides many extensions to the REST API and massive refactoring, here is a brief summary of the main features that were integrated over the last months:

• A lightweight Web-based viewer of medical images is now available as a plugin to linkimaging.
• The default SQLite engine can be replaced by an entreprise-ready database back-end based upon PostgreSQL, thanks to the powerful plugin mechanism of linkimaging.
• A DICOM Query/Retrieve client is now natively accessible from linkimaging Explorer, the embedded Web user interface of linkimaging.
• Lua scripting is getting very powerful, with full access to the REST API of linkimaging.
• For medical physicists doing quality control of imaging modalities, an extension is available for ImageJ to easily download DICOM images from linkimaging.

Very importantly, the year 2015 is marked by two awards for linkimaging:

• Sébastien, linkimaging's author, was given the Award for the Advancement of Free Software by the Free Software Foundation.
• The University Hospital of Liège won the Best eHealth Project Award for the linkimaging project from Agoria, the Belgian federation of the technology industry.

The linkimaging roadmap continues to be strongly influenced by the feedback from its growing community. Nicely, the feature set for the long-awaited stable 1.0.0 release is now getting well-defined. This release should be available in the next few months.

Thanks for your support,
Sébastien-

linkimaging is an open-source, Web-based DICOM server for healthcare and medical research. Its target audience primarily consists of the network administrators of an hospital, of the researchers in the field of biomedical imaging, as well as of the computer scientists that develop software for the automated analysis of medical images. linkimaging comes from the research of the Department of Medical Physics of the University Hospital of Liège (Belgium).

The philosophy that drives the development of linkimaging consists in proposing a different approach to Picture Archiving and Communication Systems (PACS) , that subscribes to the deconstructed PACS philosophy . Our final objective is to build a standalone component that covers the "Communication" part of the medical imaging systems. We are grateful to the I Do Imaging blog for providing an independent review of linkimaging explaining our approach.

Technically speaking, linkimaging is very different from the other free and open-source PACS software around: It is written in C++, it is lightweight (which allows it to run even on Raspberry Pi), it does not depend on any external database or framework (which makes its installation fast and trivial), and it provides a custom RESTful API . Thanks to the latter API, linkimaging can easily be driven from any computer language to automate parts of the imaging workflow that is very specific to each hospital. Thanks to its lightness, a cohort of linkimaging servers can also easily be deployed in a hospital, for instance one for each medical imaging flow.

During the past year, the most visible change to the linkimaging project was the introduction in January 2014 of its new official Web site . But many other important milestones were achieved:

• In July 2013, linkimaging 0.6.0 introduced the linkimaging Client library to call linkimaging from C++ native applications (such as viewers of medical images ).
• Since October 2013 and release 0.7.0, linkimaging can act as a DICOM Query/Retrieve server.
• In November 2013, linkimaging was nominated at Zénobe Award 2013 for social innovation.
• In April 2014, linkimaging was awarded at the Mind and Market Forum .
• In May 2014, linkimaging 0.7.5 introduced the creation of DICOM images with AJAX calls and greatly improved Store-SCU.
• In June 2014, linkimaging 0.7.6 featured the decompression of JPEG and JPEG-LS DICOM images.
• In July 2014, linkimaging 0.8.0 introduced the Plugin SDK, the Lua-based routing of images, and the support of Apple OS X. A CodeProject article explains how to turn linkimaging into a WADO server thanks to plugins.
• No later than this week, our development team is proud of having been selected by the Microsoft Innovation Center of Mons (Belgium) to take part in their SoftLab Summer Camp , a hackathon to help SMEs and public administrations to launch innovative open-source products.

Since the development roadmap is continuously shrinking, linkimaging is now heading to the major release 1.0.0 . Expect it to be ready for about Q4 2014. In the meantime, as almost no major feature will be introduced (except for Query/Retrieve SCU), please challenge linkimaging and help us finding paths for future improvements!

Thanks for your interest,
Sébastien-

We have indeed been contacted by several developers who would like to contribute to linkimaging by extending its core system so that it better fits their needs. Here are some of their various wishes:

• Add support for JPEG2000.
• Add support for WADO.
• Develop a reference implementation for standard DICOM RESTful services (WADO-RS, QIDO-RS, STOW-RS).
• Develop a reference implementation of the DICOM-RT file format for radiotherapy and nuclear medicine.
• ...

For all these reasons, we have been working hard on creating a plugin interface for linkimaging. This interface allows external developers to create shared libraries that can be loaded into the linkimaging server to meet their specific needs. Such so-called plugins can register callback functions to react to incoming REST calls. These callbacks can in turn access the linkimaging database to retrieve information about the stored DICOM files.

Here is some technical information about this development:

• Plugins must be written in C or C++. They are of course allowed to call external libraries that are more suited to their needs than DCMTK (such as GDCM, OpenJPEG...).
• The linkimaging Plugin SDK is defined in this C header . Please note that this header is licensed under the GPLv3, which notably forbids the distribution of closed-source linkimaging plugins. You can type make doc at the command-line to generate its documentation using Doxygen.
• A sample plugin is available in the linkimaging distribution.
• The path to the various plugins must be specified in the configuration file , inside the Plugins option.
• We kindly invite plugin developers to share their code in the "linkimagingContribued" GitHub repository .

Please remember that the interface and the samples are continuously evolving and getting enriched. Do not hesitate to contact us is some feature is missing, buggy or misleading.

Enjoy!
Sébastien-