Method for memorable image generation for anonymized three-dimensional medical image workflows

Patent metadata

Field	Value
Publication number	US-10722210-B2
Application number	US-201715841864-A
Country	US
Kind code	B2
Filing date	Dec 14, 2017
Priority date	Dec 14, 2017
Publication date	Jul 28, 2020
Grant date	Jul 28, 2020

Abstract

Systems and methods are provided for generating a two-dimensional image for identification of medical imaging data. An image processor acquires the medical imaging data and determines a category of the medical imaging data. A machine-learnt network identifies as a function of the category, a plurality of settings of rendering parameters that highlight one or more features the medical imaging data. The image processor renders the two-dimensional identifier image from the medical imaging data using the plurality of settings of rendering parameters and stores the medical imaging data with the two-dimensional identifier image.

First claim

I claim:

1. A method for generating a two-dimensional image for identification of medical imaging data, the method comprising:
acquiring, by an image scanner, the medical imaging data;
determining, by an image processor, a category of the medical imaging data;
identifying, using a machine-learnt network, as a function of the category, a plurality of settings of rendering parameters that highlight one or more features the medical imaging data, wherein the machine-learnt network is trained using an operator assigned memorability score generated by presenting images produced with different rendering parameters to operators and attempting by the operators to match the produced images with previously identified features, wherein images that are matched correctly are assigned a higher score;
rendering, by the image processor, a two-dimensional identifier image from the medical imaging data using the plurality of settings of rendering parameters; and
storing the medical imaging data with the two-dimensional identifier image.

2. The method of claim 1 , wherein the medical imaging data comprises computed tomography scan data.

3. The method of claim 1 , wherein the category is determined by modality, by gender of a patient, by approximate body size, and/or by a field of view of the medical imaging data.

4. The method of claim 1 , wherein the machine-learnt network is trained using a metric selected to provide a higher score for rendered image data that differ the most from a rendered canonical reference data.

5. The method of claim 4 , wherein canonical reference data comprises medical imaging data of a healthy patient.

6. The method of claim 4 , wherein canonical reference data comprises artificially generated imaging data of a standard patient.

7. The method of claim 6 , wherein the metric is selected to provide a higher score for nonempty rendered image data that differ from the rendered canonical reference data.

8. The method of claim 1 , wherein the two-dimensional identifier image is a thumbnail image.

9. A method for generating a memorable image for identification of imaging data, the method comprising:
acquiring, by an image scanner, imaging data;
identifying, using a machine-learnt network, a plurality of values of rendering parameters that highlight one or more features in the imaging data, the machine-learnt network trained using an operator assigned memorability score generated by presenting images produced with different rendering parameters to operators and allowing the operators to attempt to match the produced images with the input medical imaging data, wherein images that are matched correctly are assigned a higher score;
identifying, by an image processor, a plurality of values of reference parameters used to render a reference image;
rendering, by the image processor using the imaging data, the plurality of values of rendering parameters, and the plurality of values of reference parameters, a two-dimensional image; and
storing the imaging data with the two-dimensional image as an identifier.

10. The method of claim 9 , wherein rendering comprises:
rendering the imaging data with a combination of the plurality of values of rendering parameters and the plurality of values of reference parameters, wherein the plurality of values of rendering parameters and the plurality of values of reference parameters are combined using linear interpolation.

11. The method of claim 10 , wherein the plurality of settings of rendering parameters are weighted more heavily than the plurality of values of reference parameters.

12. The method of claim 9 , further comprising:
providing, the imaging data to an operator, when the two-dimensional image is selected.

13. The method of claim 9 , wherein the machine-learnt network is trained using a metric selected to provide a high score for rendered two-dimensional images that differ most from a rendered reference image; wherein the metric is used as a reward to train the machine-learnt network.

14. A system for generating a two-dimensional image for identification of medical imaging data, the system comprising:
a medical imaging scanner configured to acquire medical imaging data;
a machine-learnt network trained using a metric for rendered two-dimensional images that relates to a memorability score of the rendered two-dimensional images generated by presenting images produced with different rendering parameters to operators and allowing the operators to attempt to match the produced images with the input medical imaging data, the machine-learnt network configured to identify one or more values of rendering parameters for rendering a two-dimensional image from the medical imaging data;
an image processor configured to render the two-dimensional image from the medical imaging data using the one or more values of rendering parameters; and
a datastore configured to store the medical imaging data with the two-dimensional image as an identifier.

15. The system of claim 14 , wherein the medical imaging scanner is a computed tomography scanner.

16. The system of claim 14 , wherein the machine-learnt network is trained using to provide a high score for the metric for rendered two-dimensional identifier images that differ most from a rendered reference image.

17. The system of claim 14 , wherein the image processor is further configured to anonymize patient identifying data acquired with the medical imaging data.

Assignees

• Siemens Healthcare Gmbh

Inventors

• Yu Daphne

Classifications

• G16H30/00 Primary

  ICT specially adapted for the handling or processing of medical images (computed tomography A61B6/03) · CPC title

• G16H30/40

  for processing medical images, e.g. editing · CPC title

• G06F16/50

  of still image data · CPC title

• G06T2210/41

  Medical · CPC title

• G06F16/258

  Data format conversion from or to a database · CPC title