Intensity transform augmentation system and methods for use therewith

Patent metadata

Field	Value
Publication number	US-11669790-B2
Application number	US-202117336648-A
Country	US
Kind code	B2
Filing date	Jun 2, 2021
Priority date	Nov 21, 2018
Publication date	Jun 6, 2023
Grant date	Jun 6, 2023

Abstract

An intensity transform augmentation system is operable to generate a plurality of sets of augmented images by performing a set of intensity transformation functions on each of a training set of medical scans. Each of the set of intensity transformation functions are based on density properties of corresponding anatomy feature present in the training set of medical scans. A computer vision model is generated by performing a training step on the plurality of sets of augmented images, where each augmented image of a set of augmented images is assigned same output label data based on a corresponding one of the training set of medical scans. Inference data is generated by performing an inference function on a new medical scan by utilizing the computer vision model on the new medical scan. The inference data is transmitted to a client device for display via a display device.

First claim

What is claimed is:

1. An intensity transformation augmentation system, comprising:
at least one processor; and
a memory that stores operational instructions that, when executed by the at least one processor, cause the intensity transformation augmentation system to:
receive a training set of medical scans, wherein the training set of medical scans includes a first subset of medical scans that corresponds to a first anatomical region and a second subset of medical scans that corresponds to a second anatomical region;
generate a plurality of sets of augmented images, wherein each set of augmented images in the plurality of sets of augmented images is generated by performing a set of intensity transformation functions on one of the training set of medical scans, wherein each of the set of intensity transformation functions are based on density properties of corresponding one of a plurality of different anatomy features present in the training set of medical scans wherein a first proper subset of the set of intensity transformation functions is applied to medical scans of the first subset based on a first subset of the plurality of different anatomy features present in the first anatomical region, wherein a second proper subset of the set of intensity transformation functions is applied to medical scans of the second subset based a second subset of the plurality of different anatomy features present in the second anatomical region, and wherein a set difference between the first proper subset of the set of intensity transformation functions and the second proper subset of the set of intensity transformation functions is non-null;
generate a computer vision model by performing a training step on the plurality of sets of augmented images, wherein each augmented image of a set of augmented images is assigned same output label data based on a corresponding one of the training set of medical scans;
receive a new medical scan;
generate inference data by performing an inference function that utilizes the computer vision model on the new medical scan; and
transmit the inference data to a client device for display via a display device.

2. The intensity transform augmentation system of claim 1 , wherein the plurality of different anatomy features include bone, soft tissue, fat, fluid, and air.

3. The intensity transform augmentation system of claim 1 , wherein performing one of the set of intensity transformation functions on the one of the training set of the medical scans includes adjusting a contrast of image data of the one of the training set of medical scans to accentuate the corresponding one of the plurality of different anatomy features.

4. The intensity transform augmentation system of claim 1 , wherein performing one of the set of intensity transformation functions on the one of the training set of the medical scans includes adjusting a brightness of image data of the one of the training set of medical scans to accentuate the corresponding one of the plurality of different anatomy features.

5. The intensity transform augmentation system of claim 1 , wherein the training set of medical scans includes a third subset of medical scans that corresponds to a first abnormality output label and a fourth subset of medical scans that corresponds to a second abnormality output label, wherein a third proper subset of the set of intensity transformation functions is applied to medical scans of the third subset based on first density properties of the first abnormality output label, wherein a fourth proper subset of the set of intensity transformation functions is applied to medical scans of the fourth subset based on second density properties of the second abnormality output label, and wherein a set difference between the third proper subset of the set of intensity transformation functions and the fourth proper subset of the set of intensity transformation functions is non-null.

6. The intensity transform augmentation system of claim 1 , wherein the training set of medical scans includes a third subset of medical scans that corresponds to a first region of interest output label and a fourth subset of medical scans that corresponds to a second region of interest output label, wherein a third proper subset of the set of intensity transformation functions is applied to medical scans of the third subset based on a first subset of the plurality of different anatomy features that correspond to the first region of interest output label, wherein a fourth proper subset of the set of intensity transformation functions is applied to medical scans of the fourth subset based on a second subset of the plurality of different anatomy features that correspond to the second region of interest output label, and wherein a set difference between the third proper subset of the set of intensity transformation functions and the fourth proper subset of the set of intensity transformation functions is non-null.

7. The intensity transform augmentation system of claim 1 , wherein the set of intensity transformation functions is performed on image data of the training set of medical scans to produce the plurality of sets of augmented images, and wherein generating the computer vision model includes performing the training step on the image data of the training set of medical scans and the plurality of sets of augmented images.

8. The intensity transform augmentation system of claim 1 , wherein each of the set of intensity transformation functions is performed on raw sensor data of one of the training set of medical scans to generate one of a corresponding set of augmented images.

9. The intensity transform augmentation system of claim 8 , wherein the raw sensor data is in a first data format and wherein the one of the corresponding set of augmented images is in a second data format.

10. The intensity transform augmentation system of claim 9 , wherein the second data format is a JPEG image format.

11. The intensity transform augmentation system of claim 8 , wherein performing each of the set of intensity transformation functions includes calculating each one of a plurality of pixel values of the augmented image as a deterministic function of each corresponding one of a plurality of density values of the raw sensor data.

12. The intensity transform augmentation system of claim 11 , wherein the deterministic function utilized by at least one of the set of intensity transformation functions is a non-linear function of density value.

13. The intensity transform augmentation system of claim 12 , wherein the deterministic function utilized by at least one of the set of intensity transformation functions is in accordance with a cumulative distribution function of density value, and wherein the cumulative distribution function is in accordance with a normal distribution.

14. The intensity transform augmentation system of claim 1 , wherein each of the set of intensity transformation functions is in accordance with a different density window boundary pair.

15. The intensity transform augmentation system of claim 1 , wherein the operational instructions, when executed by the at least one processor, further cause the intensity transformation augmentation system to:
receive function parameter data from a client device, wherein the function parameter data corresponds to user input in response to a prompt displayed via a user interface on a display device that corresponds to the client device; and
perform at least one of the set of intensity transformation functions in accordance with the function parameter data.

16. The intensity transform augmentation system of

Assignees

• Enlitic Inc

Inventors

• Lyman Kevin
• Yao Li
• Poblenz Eric C
• Prosky Jordan
• Covington Ben
• Upton Anthony

Classifications

• G06T12/20

  Inverse problem, i.e. transformations from projection space into object space · CPC title

• G06T11/26

  Drawing of charts or graphs · CPC title

• G06T11/10

  Texturing; Colouring; Generation of textures or colours (retouching, inpainting or scratch removal G06T5/77) · CPC title

• G06N3/045

  Combinations of networks · CPC title

• G16H10/20

  for electronic clinical trials or questionnaires · CPC title