US 11,988,606 B2
Method and apparatus for multiplexed imaging of biomolecule through iterative unmixing of fluorophore signals
Jae-Byum Chang, Daejeon (KR); Young-Gyu Yoon, Daejeon (KR); Junyoung Seo, Daejeon (KR); Yeonbo Sim, Daejeon (KR); and Jee Won Kim, Daejeon (KR)
Assigned to KOREA ADVANCED INSTITUTE OF SCIENCE AND TECHNOLOGY, Daejeon (KR)
Filed by Korea Advanced Institute Of Science And Technology, Daejeon (KR)
Filed on Aug. 23, 2021, as Appl. No. 17/445,629.
Claims priority of application No. 10-2020-0106838 (KR), filed on Aug. 25, 2020; application No. 10-2020-0109519 (KR), filed on Aug. 28, 2020; application No. 10-2021-0089419 (KR), filed on Jul. 8, 2021; and application No. 10-2021-0090057 (KR), filed on Jul. 9, 2021.
Prior Publication US 2022/0065788 A1, Mar. 3, 2022
Int. Cl. G01N 21/64 (2006.01); G06T 7/00 (2017.01); G02B 21/00 (2006.01)
CPC G01N 21/6458 (2013.01) [G01N 21/6486 (2013.01); G06T 7/0014 (2013.01); G01N 2021/6419 (2013.01); G01N 2021/6421 (2013.01); G02B 21/0076 (2013.01); G06T 2207/10064 (2013.01)] 11 Claims
OG exemplary drawing
 
1. A method of an electronic device, comprising:
obtaining at least two images of at least two fluorophores marking different biomolecules, respectively; and
unmixing the at least two images into unmixed images, each of the unmixed images containing signals of a single fluorophore, while iteratively minimizing mutual information shared by the at least two images in the unmixed images,
wherein the obtaining of the at least two images of at least two fluorophores comprises obtaining two images of two fluorophores, and
wherein the unmixing of the two images comprises obtaining two unmixed images updated from the two images by applying variables for minimizing the mutual information calculated with respect to the two images, the iteration resulting in each of the unmixed images containing signals of a single fluorophore,
wherein the obtaining of the two updated unmixed images from the two images comprises processing the two images as low-resolution images having lower resolution than the two images, respectively, as in [Equation i] below;
v1=pixelBinning(u1,k)
v2=pixelBinning(u2,k)  [Equation i]
wherein the u1 and the u2 indicate the two images, respectively, the v1 and the v2 indicate the low-resolution images, respectively, and the k indicates a factor applied in order to process the two images as the low-resolution images, respectively,
calculating variables from the low-resolution images as in [Equation iii] below by using a loss function defined as in [Equation ii] below; and

OG Complex Work Unit Math
wherein the α and β the indicate the variables, respectively,
obtaining the two updated unmixed images from the two images by applying the variables and an update ratio λ to the two images as in [Equation iv] below;
u1←u1−λαu2
u2←u2−λβu1  [Equation iv]