US 12,345,637 B2
Fluid quality monitoring
Esa Hämäläinen, Nokia (FI); Tero Kesti, Tampere (FI); Teemu Heikkilä, Helsinki (FI); Vili Hätönen, Helsinki (FI); Oskari Lehto, Helsinki (FI); and Joel Pyykkö, Helsinki (FI)
Assigned to UPONOR OYJ, Vantaa (FI)
Appl. No. 18/038,111
Filed by UPONOR OYJ, Vantaa (FI)
PCT Filed Dec. 10, 2021, PCT No. PCT/FI2021/050864
§ 371(c)(1), (2) Date May 22, 2023,
PCT Pub. No. WO2022/123121, PCT Pub. Date Jun. 16, 2022.
Claims priority of application No. 20213369 (EP), filed on Dec. 11, 2020.
Prior Publication US 2024/0003808 A1, Jan. 4, 2024
Int. Cl. G01N 33/18 (2006.01); G01N 21/45 (2006.01)
CPC G01N 21/453 (2013.01) [G01N 33/18 (2013.01); G01N 2201/1296 (2013.01)] 18 Claims
OG exemplary drawing
 
1. A fluid quality measurement device, comprising:
at least one processor; and
at least one memory including computer program code;
the at least one memory and the computer program code configured to, with the at least one processor, cause the fluid quality measurement device to:
obtain a plurality of holograms, wherein each hologram in the plurality of holograms represents a microscopic object in a fluid sample;
produce a latent space representation of each hologram in the plurality of holograms using a trained autoencoder neural network;
assign each hologram in the plurality of holograms to a class in a plurality of classes based on the latent space representation of the hologram, wherein each class in the plurality of classes corresponds to a partition of the latent space; and
produce a fluid sample fingerprint based on the assignment of the plurality of holograms into the plurality of classes, wherein the fluid sample fingerprint comprises an indication of a concentration of microscopic objects in the fluid sample for each class in the plurality of classes.