Determining degree of motion using machine learning to improve medical image quality

Patent metadata

Field	Value
Publication number	US-11367179-B2
Application number	US-201916588129-A
Country	US
Kind code	B2
Filing date	Sep 30, 2019
Priority date	Sep 30, 2019
Publication date	Jun 21, 2022
Grant date	Jun 21, 2022

Abstract

Systems and techniques for determining degree of motion using machine learning to improve medical image quality are presented. In one example, a system generates, based on a convolutional neural network, motion probability data indicative of a probability distribution of a degree of motion for medical imaging data generated by a medical imaging device. The system also determines motion score data for the medical imaging data based on the motion probability data.

First claim

What is claimed is:

1. A system, comprising:
a memory that stores computer executable components; and
a processor that executes computer executable components stored in the memory, wherein the computer executable components comprise:
a machine learning component that generates, based on a convolutional neural network, motion probability data indicative of a probability distribution of a degree of motion for medical imaging data generated by a medical imaging device; and
a scoring component that determines motion score data for the medical imaging data based on the motion probability data and context data indicative of a context associated with the medical imaging data with respect to a medical condition and a patient identity, wherein the context indicates a degree of criticality of accuracy of a scan of an anatomical region associated with the patient identity by the medical imaging device, and wherein the degree of criticality of the accuracy of the scan is determined based on whether the medical condition is determined to be a life threatening medical condition associated with the patient identity.

2. The system of claim 1 , wherein the machine learning component generates the motion probability data based on an artificial recurrent neural network.

3. The system of claim 1 , wherein the scoring component calculates a normalized expected value of the probability distribution to generate the motion score data.

4. The system of claim 1 , wherein the machine learning component generates first motion probability data for a first medical image associated with the anatomical region and second motion probability data for a second medical image associated with the anatomical region.

5. The system of claim 4 , wherein the scoring component determines the motion score data based on a comparison of the first motion probability data and the second motion probability data.

6. The system of claim 4 , wherein the first medical image is associated with a first scan technique with respect to the anatomical region and the second medical image is associated with a second scan technique with respect to the anatomical region.

7. The system of claim 4 , wherein the first medical image is associated with a first contrast level and the second medical image is associated with a second contrast level.

8. The system of claim 1 , wherein the scoring component modifies initial motion score data to generate the motion score data based on the context data indicative of the context that indicates the degree of criticality of the accuracy of the scan of the anatomical region associated with the patient identity and based on a weight value associated with the degree of criticality, wherein the context relates to a scan type of the scan, a medical purpose of the scan, and a medical history associated with the patient identity, wherein, based on the context, the degree of criticality of the accuracy of the scan is determined from a group of degrees of criticality comprising a lower degree of criticality of the accuracy of the scan and a higher degree of criticality of the accuracy of the scan, wherein the higher degree of criticality is higher than the lower degree of criticality, wherein the higher degree of criticality of the accuracy of the scan is associated with the life threatening medical condition, and wherein the lower degree of criticality of the accuracy of the scan is associated with a non-life threatening medical condition.

9. The system of claim 1 , wherein the computer executable components further comprise:
an alert component that transmits an alert in response to a determination that the motion score data satisfies a defined criterion that indicates an impermissibly high degree of motion associated with the medical imaging data obtained from the scan; and
an image quality component that, in response to the alert indicating the impermissibly high degree of motion associated with the medical imaging data obtained from the scan, initiates a rescan of the anatomical region associated with the patient identity by the medical imaging device.

10. A method, comprising:
employing, by a system comprising a processor, a convolutional neural network to generate motion probability data indicative of a probability distribution of a degree of motion for medical imaging data generated by a medical imaging device; and
determining, by the system, motion score data for the medical imaging data based on the motion probability data and context data indicative of a context associated with the medical imaging data with respect to a medical condition and a patient identity, wherein the context indicates a level of criticality of accuracy of a scan of an anatomical region associated with the patient identity, and wherein the level of criticality of the accuracy of the scan is determined based on whether the medical condition is deemed to be a life threatening medical condition.

11. The method of claim 10 , wherein the employing comprises employing an artificial recurrent neural network to generate the motion probability data.

12. The method of claim 10 , further comprising:
calculating, by the system, a normalized expected value of the probability distribution.

13. The method of claim 10 , further comprising:
generating, by the system, first motion probability data for a first medical image associated with the anatomical region; and
generating, by the system, second motion probability data for a second medical image associated with the anatomical region.

14. The method of claim 13 , wherein the determining of the motion score data comprises determining the motion score data based on a comparison of the first motion probability data and the second motion probability data.

15. The method of claim 10 , wherein the determining of the motion score data comprises modifying initial motion score data to determine the motion score data based on the context data indicative of the context that indicates the level of criticality of the accuracy of the scan of the anatomical region associated with the patient identity, wherein the context relates to at least one of a scan type of the scan, a medical purpose of the scan, or a medical history associated with the patient identity, wherein the level of criticality of the accuracy of the scan is identified from a group of levels of criticality comprising a first level of criticality of the accuracy of the scan and a second level of criticality of the accuracy of the scan, wherein the second level of criticality is higher than the first level of criticality, wherein the second level of criticality of the accuracy of the scan is associated with the life threatening medical condition, and wherein the first level of criticality of the accuracy of the scan is associated with a non-life threatening medical condition.

16. The method of claim 10 , further comprising:
transmitting, by the system, an alert in response to a determination that the motion score data satisfies a defined criterion.

17. A non-transitory computer readable storage device comprising instructions that, in response to execution, cause a system comprising a processor to perform operations, comprising:
generating, using a convolutional neural network, motion probability data indicative of a probability distribution of a degree of motion for medical imaging data generated by a medical imaging device; and
determining motion score data for the medical imaging data based on the motion probability data and context data relating to a context associated with the medical imaging data with respect to a medical condition and a pati

Assignees

• Ge Prec Healthcare Llc
• Partners Healthcare System Inc
• Massachusetts Gen Hospital
• Brigham & Womens Hospital Inc
• The Bringham And Womens Hospital Inc

Inventors

• Polzin Jason
• Bizzo Bernardo
• Wright Bradley
• Kirsch John
• Schaefer Pamela

Classifications

• A61B5/055

  involving electronic [EMR] or nuclear [NMR] magnetic resonance, e.g. magnetic resonance imaging · CPC title

• G06N5/01

  Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound · CPC title

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G06N3/045

  Combinations of networks · CPC title

• G06N3/044

  Recurrent networks, e.g. Hopfield networks · CPC title