US 12,254,991 B2
Learning classifier for brain imaging modality recognition
Irina Sanchez, Barcelona (ES); Matthew Rowe, Barcelona (ES); and Santiago Puch Giner, Barcelona (ES)
Assigned to Mint Labs, Inc., Duxbury, MA (US)
Appl. No. 17/602,547
Filed by Mint Labs Inc., Boston, MA (US)
PCT Filed Apr. 13, 2020, PCT No. PCT/US2020/027986
§ 371(c)(1), (2) Date Oct. 8, 2021,
PCT Pub. No. WO2020/210826, PCT Pub. Date Oct. 15, 2020.
Claims priority of provisional application 62/832,407, filed on Apr. 11, 2019.
Prior Publication US 2022/0189014 A1, Jun. 16, 2022
Int. Cl. G16H 50/70 (2018.01); G06T 7/00 (2017.01); G06V 10/764 (2022.01); G06V 10/77 (2022.01)
CPC G16H 50/70 (2018.01) [G06T 7/0012 (2013.01); G06V 10/764 (2022.01); G06V 10/77 (2022.01); G06T 2207/20081 (2013.01); G06T 2207/20084 (2013.01); G06T 2207/30016 (2013.01)] 9 Claims
OG exemplary drawing
 
1. A method for training a model for identifying an imaging modality from a limited number of training examples, comprising:
generating, from first image data, a plurality of image vectors using a convolutional neural network;
applying a loss function to each of the plurality of image vectors to produce an intermediate dataset;
projecting the intermediate dataset in a space having lower dimensional space than the intermediate dataset;
identifying a plurality of clusters from the intermediate dataset in the space using a clustering technique; and
classifying each of the plurality of clusters into one of a plurality of imaging modalities.