SYSTEMS AND METHODS FOR SEGMENTING AN IMAGE

• US 20190333222A1
• Filed: 04/25/2019
• Published: 10/31/2019
• Est. Priority Date: 04/26/2018
• Status: Active Grant

• Current Assignee: Neuralseg Ltd.
• Original Assignee: Neuralseg Ltd.
• Inventors: Gatti, Anthony
• Granted Patent: US 11,080,857 B2
• Time in Patent Office: Days
• CPC Class Codes: G06F 18/2415 based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
• CPC Class Codes: G06T 2207/10072 Tomographic images
• CPC Class Codes: G06T 2207/10116 X-ray image
• CPC Class Codes: G06T 2207/10132 Ultrasound image
• CPC Class Codes: G06T 2207/20076 Probabilistic image processing
• CPC Class Codes: G06T 2207/20081 Training; Learning
• CPC Class Codes: G06T 2207/20084 Artificial neural networks [ANN]
• CPC Class Codes: G06T 2207/20128 Atlas-based segmentation
• CPC Class Codes: G06T 2207/30004 Biomedical image processing
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G06T 7/11 Region-based segmentation
• CPC Class Codes: G06T 7/143 involving probabilistic approaches, e.g. Markov random field [MRF] modelling
• CPC Class Codes: G06V 10/454 Integrating the filters into a hierarchical structure, e.g. convolutional neural networks [CNN]
• CPC Class Codes: G06V 10/764 using classification, e.g. of video objects
• CPC Class Codes: G06V 10/82 using neural networks
• CPC Class Codes: G06V 2201/03 Recognition of patterns in medical or anatomical images

Abstract

Methods and systems for segmenting a medical image into classes are described. A system to segment a medical image includes a processor and memory with instructions that upon execution cause the system to perform a method for segmenting the image. The method includes using initial segmentation methods to derive at least one set of probabilities of belonging to the classes for each pixel of the image. The at least one set of probabilities and the image are input into a neural network which segments the image based on both the probabilities and the image provided. This system can also use patches or sub-sections of the original image and the at least one set of probabilities as inputs to the final neural network. The patch based method enables segmentation of larger images, which usually require large amounts of time and memory to segment, and can produce a highly trained neural network.

Claims

1. A computer system for segmenting a medical image, the system comprising:

at least one processor and a memory having stored thereon instructions that, upon execution, cause the system to perform functions comprising;

inputting the medical image into at least a first segmentation method;

deriving at least one set of probabilities of belonging to at least one tissue class for each pixel of a medical image using the at least a first segmentation method;

inputting the medical image into a final neural network;

inputting the at least one set of probabilities into the final neural network; and

segmenting the medical image into the at least one tissue class based on the medical image and the at least one set of probabilities by the final neural network.

2. The system of claim 1, wherein the at least a first segmentation method includes any one or more of:

an initial neural network, a machine learning classifier, or an atlas-based segmentation algorithm.

3. The system of claim 1, wherein the medical image is input into at least one of the at least a first segmentation method and the final neural network as sub-sections of the medical image, the method further comprising:

deriving at least one set of probabilities for each sub-section of the medical image; and

combining the probabilities from the sub-sections.

4. The system of claim 1, wherein the functions further comprise pre-processing the medical image.

5. The system of claim 1, wherein the at least one set of probabilities is derived from a lower resolution iteration of the medical image.

6. The system of claim 1, wherein the at least one set of probabilities is derived from at least two iterations of the medical image.

7. The system of claim 1, wherein the medical image is any one of:

a magnetic resonance imaging image, a computed tomography image, an ultrasound image, an x-ray image, or a positron emission tomography image.

8. A method of segmenting an image, the method comprising:

deriving at least one set of probabilities of belonging to m classes, where m is any positive integer, for each pixel of an image using at least one segmentation method;

inputting the image into a final neural network;

inputting the at least one set of probabilities into the final neural network; and

segmenting the image into the m classes based on the image and the at least one set of probabilities by the final neural network.

9. The method of claim 8, wherein the at least a first segmentation method includes any one or more of:

an initial neural network, a machine learning classifier, or an atlas-based segmentation algorithm.

10. The method of claim 8, wherein the image is input into at least one of the at least a first segmentation method and the final neural network as sub-sections of the image, the method further comprising:

deriving at least one set of probabilities for each sub-section of the image; and

combining the probabilities from the sub-sections.

11. The method of claim 8, wherein the functions further comprise pre-processing the image.

12. The method of claim 8, wherein the at least one set of probabilities is derived from a lower resolution iteration of the image.

13. The method of claim 8, wherein the at least one set of probabilities is derived from at least two iterations of the image.

14. The method of claim 8, wherein the image is any one of:

a magnetic resonance imaging image, a computed tomography image, an ultrasound image, an x-ray image, or a positron emission tomography image.

15. A system for segmenting an image, the system comprising:

at least one processor and a memory having stored thereon instructions that, upon execution, cause the system to perform function comprising;

deriving at least one set of probabilities of belonging to m classes where m is any integer, for each pixel of an image using at least a first segmentation method;

inputting the image into a final neural network;

inputting the at least one set of probabilities into the final neural network; and

segmenting the image into the m classes based on the image and the at least one set of probabilities by the final neural network.

16. The system of claim 15, wherein the at least a first segmentation method includes at least one method selected from the group consisting of:

an initial neural network, a machine learning classifier, or an atlas-based segmentation algorithm.

17. The system of claim 15, wherein the image is input into at least one of the at least a first segmentation method and the final neural network as sub-sections of the image, the method further comprising:

deriving at least one set of probabilities for each sub-section of the image; and

combining the probabilities from the sub-sections.

18. The system of claim 15, wherein the functions further comprise pre-processing the image.

19. The system of claim 15, wherein the at least one set of probabilities is derived from a lower resolution iteration of the image.

20. The system of claim 15, wherein the at least one set of probabilities is derived from at least two iterations of the image.