I hope there will be time to discuss Terminologies improvements, @muratmaga is joining as well. It would be great to agree on how to add creating custom terminology (see here).
Module Improvements: Discussed potential enhancements to the Terminologies module. See the discussion here: Slicer Issue #6975.
Terminology, Color Table, and Segment Relationship:
There is a significant disparity between the vast number of available terminology terms and the limited number of segments. The color table serves as a bridge:
Analogy: Terminology represents all available paint colors, the color table is the palette, and segments are the canvas.
Handling Missing Terms:
It is not problematic if a segment lacks a corresponding term in the terminology. A streamlined editing approach could involve direct color table modifications. This would require an improved file format for color tables (details below).
Idea: Develop a new Terminologies extension to connect with an online repository of terminologies (e.g., on GitHub).
Color Table File Format Enhancements:
Transition to CSV format, ensuring compatibility with scene load and save operations.
Include all essential terminology information: name, code, coding scheme designator, category, modifier, and optional anatomical location.
Allow for JSON properties to be optional, omitting unused entries.
Displaying Color Tables in Terminology Contexts:
Limit to discrete options and filter based on specific attributes (e.g., “compatible”: adequate metadata).
Creating a “Short List” Color Table:
Enable users to generate a new color table based on current segments, providing a customizable starting point.
MorphoDepo (MorphoCloud) Requirements:
Each user receives a unique color table tied to their segmentation.
Hierarchical structure required, such as categories (e.g., “skeletal system / spine”) to create a more organized tree structure, replacing the current flat list.
Next Steps:
Enhance color tables to store metadata effectively.
Integrate color tables within the terminology selector for streamlined user experience.
Slicer & Trame
Trame Integration:
Discussed the integration of trame with Slicer to support backend GPU rendering.
Render window content is converted into a compressed JPEG image, which is transmitted to the front end as a web application.
Transitioning to in-render window widgets would minimize dependencies on Qt.
Refactoring will be required to enable segmentation editor effects to function independently from Qt.
On-going development of trame-based Slicer should be released open-source in the first quarter of 2025. For questions, reach out to @finetjul, @Thibault_Pelletier and @jcfr