IP Library › Granted Patent US 12694645
Granted Patent B2
US 12694645 · App. 18/063,669 · Granted Jul 28, 2026

Systems and methods for automatically detecting substances in medical imaging

Inventors: Brian Fouts (Morgan Hill, CA); Cole Kincaid Hunter (Santa Clara, CA); Aric Josef Beikmann (San Jose, CA); John Jingan Tang (San Jose, CA); Wenjing Li (Sunnyvale, CA); Davin Trey Fish (Provo, UT); Brady Lewis Woolford (Mapleton, UT)
Assignee: Stryker Corporation
G06V10/764
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12694645
App. No.
18/063,669
Granted
Jul 28, 2026
Kind
B2
Abstract

A method for training a machine learning model to detect substances that compromise medical imaging clarity includes receiving imaging of tissue associated with a medical procedure type; receiving imaging of one or more substances that can affect clarity of imaging associated with the medical procedure type; combining at least a portion of the imaging of the tissue and at least a portion of the imaging of the one or more substances to generate machine learning training image data in which at least a portion of the tissue is at least partially obscured by the one or more substances; and training, with the training image data, a machine learning model to detect the one or more substances in imaging generated during a medical procedure of the medical procedure type.

Claims (28)

1 . A method for training a machine learning model to detect substances that compromise medical imaging clarity, the method comprising:

receiving imaging of tissue associated with a medical procedure type;

receiving imaging of one or more substances that can affect clarity of imaging associated with the medical procedure type, wherein the one or more substances in the imaging of the one or more substances are not inside a body of a subject;

combining at least a portion of the imaging of the tissue and at least a portion of the imaging of the one or more substances to generate machine learning training image data in which at least a portion of the tissue is at least partially obscured by the one or more substances; and

training, with the training image data, a machine learning model to detect the one or more substances in imaging generated during a medical procedure of the medical procedure type.

2 . The method of claim 1 , wherein the imaging of the tissue and the imaging of the one or more substances were captured by the same type of imaging device.

3 . The method of claim 1 , wherein the imaging of the tissue and the imaging of the one or more substances were captured by one or more endoscopic imagers.

4 . The method of claim 1 , wherein the imaging of the one or more substances was generated using a fixture through which the one or more substances were directed.

5 . The method of claim 4 , wherein combining at least a portion of the imaging of the tissue and at least a portion of the imaging of the one or more substances comprises isolating the one or more substances in the at least a portion of the imaging of the one or more substances using a clear frame of the fixture.

6 . The method of claim 1 , wherein the imaging of tissue associated with a medical procedure type is free of the one or more substances.

7 . The method of claim 1 , wherein the one or more substances comprise smoke, blood, debris, or bubbles.

8 . The method of claim 7 , wherein the blood is synthetic blood or blood modified with anticoagulant.

9 . The method of claim 1 , wherein combining at least a portion of the imaging of the tissue and at least a portion of the imaging of the one or more substances comprises isolating portions corresponding to the one or more substances in the at least a portion of the imaging of the one or more substances, inverting colors of the isolated portion, and subtracting the inverted colors of the isolated portions from the at least a portion of the imaging of the tissue.

10 . The method of claim 1 , wherein combining at least a portion of the imaging of the tissue and at least a portion of the imaging of the one or more substances comprises at least one alignment, rotation, or translation step.

11 . The method of claim 1 , wherein combining at least a portion of the imaging of the tissue and at least a portion of the imaging of the one or more substances comprises extracting respective circles from the at least a portion of the imaging of the tissue and the at least a portion of the imaging of the one or more substances, determining a resizing for matching the respective circles, and applying the resizing to at least one of the at least a portion of the imaging of the tissue and the at least a portion of the imaging of the one or more substances.

12 . The method of claim 1 , wherein the at least a portion of the imaging of the tissue is combined with the at least a portion of the imaging of the one or more substances using a weighted average.

13 . A method for detecting substances that compromise medical imaging clarity, the method comprising:

receiving imaging generated during a medical procedure of a medical procedure type; and

detecting one or more substances that compromise clarity in the imaging generated during the medical procedure using a machine learning model, wherein the machine learning model was trained on training image data generated by combining at least a portion of imaging of tissue associated with the medical procedure type and at least a portion of imaging of the one or more substances, wherein the one or more substances in the imaging of the one or more substances are not inside a body of a subject, such that at least a portion of the tissue is at least partially obscured by the one or more substances.

14 . The method of claim 13 , wherein the imaging generated during a medical procedure, the imaging of the tissue, and the imaging of the one or more substances were captured by the same type of imaging device.

15 . The method of claim 13 , wherein the imaging generated during a medical procedure, the imaging of the tissue, and the imaging of the one or more substances were captured by one or more endoscopic imagers.

16 . The method of claim 13 , wherein the imaging of the one or more substances was generated using a fixture through which the one or more substances were directed.

17 . The method of claim 16 , wherein the at least a portion of the imaging of the tissue and the at least a portion of the imaging of the one or more substances were combined by isolating the one or more substances in the at least a portion of the imaging of the one or more substances using a clear frame of the fixture.

18 . The method of claim 13 , wherein the one or more substances comprise smoke, blood, debris, or bubbles.

19 . The method of claim 13 , wherein the training image data comprises multiple combinations of the at least a portion of the imaging of the tissue and the at least a portion of the imaging of the one or more substances, each combination differing in at least one of rotation and translation of the at least a portion of the imaging of the one or more substances.

20 . A system for detecting substances that compromise medical imaging clarity, the system comprising one or more processors, memory, and one or more programs stored in the memory for execution by the one or more processors for:

receiving imaging generated during a medical procedure of a medical procedure type; and

detecting one or more substances that compromise clarity in the imaging generated during the medical procedure using a machine learning model, wherein the machine learning model was trained on training image data generated by combining at least a portion of imaging of tissue associated with the medical procedure type and at least a portion of imaging of the one or more substances, wherein the one or more substances in the imaging of the one or more substances are not inside a body of a subject, such that at least a portion of the tissue is at least partially obscured by the one or more substances.