GTCreator: a flexible annotation tool for image-based datasets

Area of Science:

• Medical Imaging
• Health Informatics
• Computer-Aided Diagnosis

Background:

• Evaluating health decision support systems, particularly polyp detection in colonoscopy, requires extensive image annotation.
• Current annotation tools lack flexibility and ease of use, hindering efficient methodology evaluation.

Purpose of the Study:

• To introduce GTCreator, a novel annotation tool designed to improve the flexibility and efficiency of image and text annotation for image-based datasets.
• To address the time-consuming nature of methodology evaluation for health decision support systems.

Main Methods:

• GTCreator offers core annotation functionalities with extended capabilities, including simultaneous multi-annotator support.
• Features enhanced dataset browsing and efficient annotation transfer to accelerate the annotation process for large datasets.

Main Results:

• GTCreator provides fast and precise annotation of image datasets.
• It is the only tool offering comprehensive annotation editing and browsing functionalities.

Conclusions:

• GTCreator proves efficient for annotating large image datasets, significantly aiding in the evaluation of medical imaging analysis methods.
• The tool shows potential for broader applications in method evaluation, including experimental setup and results analysis.