Graph neural network–derived SLIC tile features for human synovial histology 02

This dataset contains graph neural network (GNN)–derived feature embeddings for SLIC superpixel tiles generated from human synovial histology images. Each file corresponds to a single case and provides 64-dimensional feature vectors for each tile, representing learned spatial and contextual tissue information. This record corresponds to Part 2 of the GNN-derived feature collection.