An ordered record of events.
Chronicle is a system to manage a directed acyclic graph of data items.
The items are represented as DICOM object instances stored as documents in CouchDB.
site: a web app (html5/js/css) to be hosted from CouchDB for interacting with the data.
bin: python utility scripts to transfer local (DICOM) data to a CouchDB instance.
Use https://github.com/pieper/couchSite to upload design documents and site (html/css/javascript)
Install a python environment with the prerequisites above (has been tested on mac and linux).
Install Apache couchdb
Clone the chronicle respository: git clone git://github.com/pieper/Chronicle
Get the couchSite utility: git clone git://github.com/pieper/couchSite
Install chronicle into couchdb: cd Chronicle; ../couchSite/couchSite.py . chronicle
At this point you should have an empty chronicle database and the web app. The next step is to install some DICOM data into the database, which can be done as follows from the chronicle directory: ./bin/record.py <path to dicom data>
Note that only pixel formats (transfer syntaxes) supported by pydicom can be used, so some compressed images cannot be loaded currently.
The current Chronicle prototype includes functionality for viewing control-point curves on a per-slice basis and for 3D rendering of these structures. At this point there is no way to create these using the user interface and there are no publicly available datasets that include these annotations.
This is a pre-release project and is subject to change at any time with no assurance of backward compatibility.
This work is supported by NIH National Cancer Institute (NCI) through award U24 CA180918 (QIICR: Quantitative Image Informatics for Cancer Research) and the National Institute of Biomedical Imaging and Bioengineering (NIBIB) through awards P41 EB015902 (NAC: Neuroimage Analysis Center) and U54 EB005149 (NA-MIC: National Alliance for Medical Image Computing). Additional support provided by Novartis AG.