USPTO Patent US12586184B2: Pathology Pattern Analysis via Graph Deep Learning

ChangeBridge: Patent Grants - AI & Computing (G06N)

Published March 24th, 2026

Detected March 25th, 2026

Summary

The USPTO has granted patent US12586184B2 to Seoul National University R&DB Foundation for a method and apparatus analyzing pathology patterns of whole-slide images using graph deep learning. The patent describes techniques for compressing images, analyzing them with graph neural networks, and extracting diagnostic information.

View original document View source feed page

What changed

The United States Patent and Trademark Office (USPTO) has granted patent US12586184B2, titled 'Pathology pattern analysis via graph deep learning,' to Seoul National University R&DB Foundation. The patent details a method and apparatus for analyzing whole-slide images (WSIs) using graph deep learning, involving image compression into a superpatch graph, analysis via graph neural networks (GNNs) to calculate node and edge contributions, biomarker acquisition through classification and clustering of graph features, and extraction of diagnostic information. The patent was filed on March 7, 2023, and granted on March 24, 2026.

This patent grant represents a new technological development in AI-driven pathology analysis. While not a regulatory rule imposing obligations, it signifies innovation in the field that may influence future medical device development and diagnostic tools. Compliance officers in the pharmaceutical and medical device sectors should be aware of this patented technology as it may impact product development roadmaps, intellectual property considerations, and potential collaborations or licensing opportunities.

Source document (simplified)

← USPTO Patent Grants

Methods and apparatus for analyzing pathology patterns of whole-slide images based on graph deep learning

Grant US12586184B2 Kind: B2 Mar 24, 2026

Assignee

SEOUL NATIONAL UNIVERSITY R&DB FOUNDATION

Inventors

Sunghoon Kwon, Yongju Lee, Kyoungseob Shin, Kyung Chul Moon, Jeong Hwan Park, Sohee Oh

Abstract

The present invention relates to a method and apparatus for analyzing pathology patterns of whole-slide images based on graph deep learning, which may include: a whole-slide image (WSI) compression step of compressing WSI into a superpatch graph; a graph neural networks (GNN) analysis step of embedding node features and context features into the superpatch graph through a GNN model and calculating contributions for each node and edge; a biomarker acquisition step of classifying and grouping the superpatch graph according to the contributions for each node, connecting the classified and grouped superpatch graph in units of groups to generate connected graphs, normalizing and clustering features of the connected graphs, and acquiring environmental graph biomarkers for each group; and a diagnostic information extraction step of extracting and providing diagnostic information on the WSI based on the environmental graph biomarker for each group.

CPC Classifications

G06T 7/0012 G06T 7/11 G06T 2207/20084 G06T 2207/20072 G06T 2207/30004 G06T 5/40 G06T 2210/41 G06T 2207/20081 G06T 2207/20021 G06T 7/10 G06T 7/162 G06T 2207/20112 G06T 7/248 G06T 2207/30024 G06T 2207/10056 G06T 7/337 G06T 5/60 G06V 10/44 G06V 10/762 G06V 10/764 G06V 10/454 G06V 10/7635 G06V 10/82 G06V 20/69 G06V 10/70 G06V 2201/03 G06V 20/695 G06V 10/26 G06V 10/40 G06V 10/42 G06V 10/426 G06V 20/46 G06V 40/171 G06V 20/693 G06V 10/84 G06V 10/86 G06V 10/421 G06V 20/49 G06V 20/00 G06V 2201/122 G06V 40/382 G06V 10/77 G06V 10/7715 G06V 10/771 G06V 10/806 G06V 10/50 G06V 20/698 G06V 40/1347 G16H 50/20 G16H 50/30 G16H 30/40 G16H 50/00 G16H 70/60 G06N 3/08 G06N 20/00 G06N 5/04 G06N 3/082 G06N 5/045 G06N 7/01 G06N 3/02 G06N 3/0464 A61B 8/5207 A61B 8/5223 A61B 8/52 A61B 90/20