An easy-to-use and accessible labeling platform. The platform can be used to assess human observer labels for image quality assessment or to perform a blinded reading or a qualitative evaluation.
The platform is built up as a client-server architecture with a lightweight client (HTML web page frontend) and processing and encrypted storing on server side, as illustrated in the Figure below.
Users can register to the platform and provides some background information to allow the assignment for a specific test-case. Afterwards they automatically receive an invitation to participate in a labeling when a new labeling test-case is generated. The users login to the platform via a web browser.
Test-case
consists of:
and the displayed web page front end contains
(a) one displayed image, i.e. independent labeling
(b) 1 to N images in parallel, i.e. in relation to each other
(c) a reference image and 1 to N images in parallel
User types
- admin: generate and manage test-cases, manage users, analyze and export labels, maintain image database
- user: access labeling test-cases
Images
- images can be in DICOM, MHD or GIPL format
- standard viewing capabilities: zooming, rotating, brightness and contrast adjustment
- viewing is realized via SimpleITK: https://github.com/SimpleITK/SimpleITK
The labeling platform is publicly available under Apache 2.0 license: https://github.com/thomaskuestner/LabelingPlatform
Please cite the paper, if you use it in a scientific publication.
Utility/Constant.cs
Web.config (line 9): <add name="MySqlConnection" connectionString="Server=SERVERNAME;Database=labelingframework;Uid=SQL_ADMIN;Pwd=SQL_ADMIN_PWD;" />
Styles/Site.css
LabelingFramework.csproj
and compilexbuild /p:Configuration=Release LabelingFramework.csproj
MySQL/default_database.sql
Structure of labeling web page frontend
More information on the usage and the structure can be viewed directly in the labeling platform: https://github.com/thomaskuestner/LabelingPlatform/usage_help.html
or
close testcase
help
tutorial page (additional testcase information)
report a problem
reset image (zoom, brightness, contrast)
link images