Skip to content

v1.0.0 – Rule-based Visual Attention Core

Choose a tag to compare

@its-spark-dev its-spark-dev released this 28 Jan 07:27
· 1 commit to main since this release

This release marks the initial stable version (v1.0) of the rule-based visual attention core.

The core provides:

  • A deterministic, stateless Feature abstraction for visual attention.
  • Multiple example attention features (e.g., center bias, contrast, edge-based saliency).
  • A simple and extensible feature fusion mechanism.
  • Clear separation between core computation and application-level concerns.

This repository intentionally excludes:

  • Image I/O
  • Visualization or heatmap rendering
  • Web or CLI applications
  • Machine learning models or training pipelines

Future work (demos, visualization, and applications) will be developed in separate repositories on top of this core.