Nora is a web-based framework for medical image analysis.
It has been developed to bridge the gap between research and clinic, and to
boost medical imaging research to the next level.
It provides a high-level web-interface accessible from any webbrowser to visualize, organise, process and share data in a very customizable way.
Depending on your needs, your Nora instance can run as a web-service in the cloud or as a local installation at your institution.
The Nora image viewer runs in any web-browser, without installation or update issues.
It provides features beyond the standard, such as real-time image resclicing (MPR), overlays, ROIs,
3D surface rendering and connectome/fiber viewer.
Besides medical image file formats like DICOM, NIFTI and BRUKER, the viewer has functionalities to show other common formats like json, jpeg, png, pdf.
With this, you will always have all relevant data at one glance.
The viewer is highly customizable: For different projects, it can be configured with preset schemes to automatically load specific datasets with specific settings.
The viewer also works as a standalone offline tool: Without installation, simply drag and drop your local files into the browser and use all of the useful features.
Manage, visualise, process and analyse your images on a web-based multi-user platform.
Include data from various inhomogeneous sources, no matter if file-based, spread over multiple harddrives, or imported from a PACS system.
In the webinterface, you will always have all data at one glance.
For image processing, Nora provides a unified interface.
Use existing toolboxes such as SPM, FSL or FreeSurfer and others, without dealing with their individual function calls,
or implement your own algorithms with python, matlab or as command line tool.
You can simply define your batch pipelines in the webinterface, and start processing hundreds of patients on a parallel grid engine with a single mouseclick.
Finally, retrieve a results table with calculated quantitative parameters in selected regions of interest chosen from built-in atlases or segmented manually.
With Nora, you can effectively join all your data for a specific project.
No matter how your data is distributed, simply add your data folders to your project, Nora will gather all information in a database.
The Nora data management can deal with any data folder organisation, as long as some ID can be extracted from the directory names.
This way, you can join data from many different locations and store calculated results in another folder structure or location, without loosing track.
You can also add remote data from mounted network drives or cloud storage.
The Nora database only holds meta information, the data itself always remains where it is.
The system also provides a DICOM interface to exchange data with other PACS systems or imaging devices.
No matter how your data is distributed, in the web-interface everything will show up well organised at one glance, searchable, browsable and sharable to colleagues working on this project.
Here, you can also set up your SMARTLOADERS to define which images should be loaded in which configuration when selecting a patient.
A new Patient, sent from a PACS, uploaded via the webinterface or mounted as a new network drive will automatically pop up in the list.
You can set tags, add metadata, browse, sort, search, define groups and and start the data processing from here.
Nora provides a high level interface for existing imaging processing toolboxes, or your own code.
No matter if you want to process data with your matlab code, command line tools or toolboxes like SPM, FSL or FreeSurfer:
Don't care about the specific function calls of the different programming languages and toolboxes.
Define your calculation pipeline via the webinterface and process thousands of datasets with a single mouseclick.
A grid engine will take care for parallel computing of your jobs, with full control over optimal workload, job status, progress and errors.
In many cases, data from clinical trials is evaluated based on statistical measures.
These can be general global parameters, such as clinical scores, or image-derived parameters from functional MRI, diffusion or perfusion imaging, or any other quantitative measure.
In many cases, these values are to be compared in selected regions of interest.
With a single mouseclick, you can aggregate these values for huge clinical trials and retrieve a table of results, joined with any other metadata.
Nora allows for interactive reading, labeling and tagging in a multi-user environment.
You want to investigate whether prostate cancer can better be assessed using a combination of PET/CT or DCE MRI?
Or you want to know if your automated segmentation algorithm really can support visual diagnosis of lung cancer?
Set up a project and define an interactive form with click, check, and combo-boxes.
Invite your colleagues worldwide to read and rate the cases in a randomized manner and / or in multiple stages showing the data in different configurations.
With this standardisation, you can get results with great statistical power.
This is not only a powerful tool for randomized surveys, but also to perform quality checks and to generate ground truth datasets with labels, annotations and regions of interest.
Invite readers from your intranet or from all over the world to participate in your reading study.
No matter if they sit in their office in New York, or at the airport in Shanghai,
they can walk through the cases from anywhere in the world, whenever they want.
The system is easily scalable: Once your reading is set up, it doesn't matter how many readers access the system.
With this, your statistical significance will increase enormously, and you can create very reliable ground truth datasets.
You can always follow the current status of the reading.
Once all readers have done their job, download the results table and publish your high impact paper.
Applications with fully automated processing pipelines can provide very valuable support for clinical diagnosis.
Especially the emerging field of machine learning methods, based on the combination of multimodal data is regarded as an important tool for improved personalised healthcare.
With Nora, scientists, software engineers and manufacturers can provide their applications to a large community in order to test and improve their CAD methods.
Clinicians all over the world, not only from large clinical centers but also from smaller hospitals can gain access to cutting edge research
technologies.
Contact:
NORA Medical Imaging Platform Project
Dr. Elias Kellner
Dr. Marco Reisert
University Medical Center Freiburg
Department of Radiology
Medical Physics
Killianstr. 5a
79106 Freiburg, Germany
+49 761 270 93860
elias.kellner[at]uniklinik-freiburg.de
marco.reisert[at]uniklinik-freiburg.de