Supplementary Minimal Dataset for "Spatial–Attention Collaborative Framework for Controllable Text-Guided Image Editing"

This upload provides a minimal supplementary dataset and configuration files to support the reproducibility of the experiments in the manuscript “Spatial–Attention Collaborative Framework for Controllable Text-Guided Image Editing” (Electronics, under review). The package includes JSON annotation files and scripts for generating evaluation subsets. No original images from third-party datasets (e.g., Stanford Cars, MS-COCO) are redistributed. Users should download the original images from the official sources according to the provided instructions.