US 12,462,406 B2
Systems and methods for slide image alignment
Alexander Anderson, Tampa, FL (US); Mark Robertson-Tessi, Lutz, FL (US); and Chandler Gatenbee, Tampa, FL (US)
Assigned to H. LEE MOFFITT CANCER CENTER AND RESEARCH INSTITUTE, INC., Tampa, FL (US)
Appl. No. 17/609,922
Filed by H. LEE MOFFITT CANCER CENTER AND RESEARCH INSTITUTE, INC., Tampa, FL (US)
PCT Filed May 11, 2020, PCT No. PCT/US2020/032321
§ 371(c)(1), (2) Date Nov. 9, 2021,
PCT Pub. No. WO2020/227710, PCT Pub. Date Nov. 12, 2020.
Claims priority of provisional application 62/845,585, filed on May 9, 2019.
Prior Publication US 2022/0237810 A1, Jul. 28, 2022
Int. Cl. G06T 7/33 (2017.01); G06T 3/14 (2024.01); G06T 3/4038 (2024.01); G06V 10/24 (2022.01); G06V 10/74 (2022.01); G06V 10/75 (2022.01); G06V 10/762 (2022.01); G06V 20/69 (2022.01)
CPC G06T 7/337 (2017.01) [G06T 3/14 (2024.01); G06T 3/4038 (2013.01); G06V 10/24 (2022.01); G06V 10/751 (2022.01); G06V 10/761 (2022.01); G06V 10/7625 (2022.01); G06V 20/695 (2022.01); G06T 2207/30024 (2013.01)] 19 Claims
OG exemplary drawing
 
1. A computer-implemented method for slide image registration, comprising:
receiving a plurality of slide images;
detecting a plurality of features contained in the slide images;
comparing a plurality of pairs of the slide images, wherein the comparison uses the detected features;
creating a distance matrix that reflects a respective difference between each of the pairs of the slide images;
creating a graph by connecting each of the slide images to its most similar slide image in feature space, wherein each slide image is a vertex of the graph, and each connected slide image pair defines an edge of the graph;
detecting a plurality of graph components using a community detection operation, wherein each of the graph components comprises one or more of the slide images;
aligning the slide images within each of the graph components by aligning two or more slide images within each graph component using at least one feature shared by the two or more slide images within the respective graph component;
refining the alignment of the two or more slide images by using an optical flow technique to estimate a displacement of each slide image from a previous slide image; and
generating a composite image or series based, at least in part, on the aligned graph components.