Granted Patent B2

US 12,406,753 · App. 18/834,427 · Granted Sep 2, 2025

Inventors: Yao-Jen Chang (Princeton, NJ); Vivek Singh (Princeton, NJ); Boris Mailhe (Plainsboro, NJ); Benjamin S. Pollack (Jersey City, NJ); Ankur Kapoor (Plainsboro, NJ)

Assignee: Siemens Healthcare Diagnostics Inc.

G16H10/40

Abstract

Claims (34)

1. A method of synthesizing an image of a tube assembly, comprising:

capturing an image of a tube assembly, the capturing generating a captured image;

decomposing the captured image into a plurality of features in latent space using a trained image decomposition model;

manipulating one or more of the features in the latent space into manipulated features; and

generating a synthesized tube assembly image with at least one of the manipulated features using a trained image composition model.

2. The method of claim 1 , further comprising training an image decomposition model and an image composition model using at least an image of a first tube assembly and an image of a second tube assembly, wherein there are one or more known variations between the image of the first tube assembly and the image of the second tube assembly, the training producing the trained mage image decomposition model and the trained image composition model.

3. The method of claim 2 , wherein the training the decomposition model is at least partially based on a reconstruction loss between synthesized images and input images.

4. The method of claim 1 , wherein the tube assembly comprises a tube and a cap and at least one of the plurality of features is tube geometry, tube color, tube surface property, cap geometry, cap color, or lighting condition.

5. The method of claim 1 , wherein the tube assembly comprises a tube and at least one of the plurality of features is tube geometry, tube material, sample fluid in the tube, label affixed to the tube, or lighting condition of the tube.

6. A method of synthesizing images of tube assemblies, comprising:

constructing an image decomposition model configured to receive input images of tube assemblies and to decompose the input images into a plurality of features in a latent space;

constructing an image composition model configured to compose synthetic tube assembly images based on the plurality of features in the latent space; and

training the image decomposition model and the image composition model using at least an image of a first tube assembly and an image of a second tube assembly with one or more known variations between the image of the first tube assembly and the image of the second tube assembly, wherein the training produces a trained image decomposition model and a trained image composition model.

7. The method of claim 6 , further comprising:

capturing an image of a tube assembly, the capturing generating a captured image;

decomposing the captured image into a plurality of decomposed features in the latent space using the trained image decomposition model;

manipulating one or more of the decomposed features in the latent space into manipulated features; and

generating a synthesized tube assembly image with at least one of the manipulated features using the trained image composition model.

8. The method of claim 6 , wherein the training the decomposition model is at least partially based on a reconstruction loss between synthesized images and input images.

9. The method of claim 6 , wherein the image composition model is trained based on a reconstruction loss between synthesized images and input images.

10. The method of claim 6 , wherein at least one of the features in the latent space is generated using computer-generated imagery.

11. The method of claim 6 , wherein images of at least two of the tube assemblies differ from each other with one or more controlled attributes.

12. The method of claim 6 , wherein images of at least two tube assemblies share one or more attributes in common.

13. The method of claim 6 , wherein at least one of the image decomposition model or the image composition model is trained based on a reconstruction loss without the images of the tube assemblies.

14. The method of claim 6 , wherein the tube assemblies comprise a tube and a cap and at least one of the plurality of features in the latent space is tube geometry, tube color, tube surface property, cap color, cap geometry, or lighting condition.

15. The method of claim 6 , wherein the tube assemblies comprise a tube and at least one of the plurality of features in the latent space is tube geometry, tube material, sample fluid in the tube, label affixed to the tube, or lighting condition of the tube.

16. The method of claim 6 , wherein one or more features in the latent space include a fluid property in a tube assembly.

17. A diagnostic laboratory system, comprising:

an image decomposition model configured to receive input images of tube assemblies and to decompose the input images into a plurality of features in a latent space; and

an image composition model configured to compose synthetic tube assembly images based on the plurality of features in the latent space,

wherein the image decomposition model and the image composition model are trained using at least an image of a first tube assembly and an image of a second tube assembly with one or more known variations between the image of the first tube assembly and the image of the second tube assembly.

18. The diagnostic laboratory system of claim 17 , wherein the image decomposition model is trained at least partially based on a reconstruction loss between synthesized images and input images.

19. The diagnostic laboratory system of claim 17 , wherein the image composition model is trained based on a reconstruction loss between synthesized images and input images.

20. The diagnostic laboratory system of claim 17 , wherein the tube assemblies comprise a tube and a cap and at least one of the plurality of features in the latent space is tube geometry, tube color, tube surface property, cap color, cap geometry, or lighting condition.

Assignments (3)

Continuity (2)

References Cited (56)