← Back to article | SpatialFinder: a human-in-the-loop vision-language framework for prioritizing high-value regions in spatial transcriptomics