AUTOMATED ANATOMICAL LABELING BY MULTI-CONTRAST DIFFEOMORPHIC PROBABILITY FUSION

• US 20160343127A1
• Filed: 01/20/2015
• Published: 11/24/2016
• Est. Priority Date: 01/17/2014
• Status: Active Grant

• Current Assignee: Johns Hopkins University
• Original Assignee: Johns Hopkins University
• Inventors: Tang, Xiaoying, Miller, Michael I., Mori, Susumu
• Granted Patent: US 10,535,133 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: A61B 5/055 involving electronic [EMR] or nuclear [NMR] magnetic resonance, e.g. magnetic resonance imaging
• CPC Class Codes: A61B 6/032 Transmission computed tomography [CT]
• CPC Class Codes: A61B 6/037 Emission tomography
• CPC Class Codes: A61B 6/5211 involving processing of medical diagnostic data
• CPC Class Codes: A61B 8/5215 involving processing of medical diagnostic data
• CPC Class Codes: G06F 18/24 Classification techniques
• CPC Class Codes: G06T 2207/10081 Computed x-ray tomography [CT]
• CPC Class Codes: G06T 2207/10088 Magnetic resonance imaging [MRI]
• CPC Class Codes: G06T 2207/10104 Positron emission tomography [PET]
• CPC Class Codes: G06T 2207/20021 Dividing image into blocks, subimages or windows
• CPC Class Codes: G06T 2207/20076 Probabilistic image processing
• CPC Class Codes: G06T 2207/20128 Atlas-based segmentation
• CPC Class Codes: G06T 2207/30016 Brain
• 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: G06T 7/174 involving the use of two or more images
• CPC Class Codes: G06T 7/33 using feature-based methods

Abstract

A computer-implemented method, system and non-transitory computer readable storage medium for classifying a region of interest of a subject, including receiving imaging data comprising at least one image element, the imaging data comprising the region of interest of the subject; providing a plurality of atlases, each of the plurality of atlases having a candidate region that corresponds to the region of interest of the imaging data, each of the plurality of atlases having at least one image element with associated location and property information; co-registering the plurality of atlases to the imaging data, using at least one processor; assigning a probability to generate a labeling parameter for the region of interest, the probability being associated with each atlas; and classifying the region of interest of the subject based on the assigning.

Claims

2-27. -27. (canceled)

28. A system for classifying a region of interest of a subject, the system comprising:

a memory storing computer-executable instructions; and

a processor that is coupled to the memory and that is configured to execute the computer-executable instructions to perform;

receiving imaging data comprising at least one image element, the imaging data comprising the region of interest of the subject;

providing a plurality of atlases, each of the plurality of atlases having a candidate region that corresponds to the region of interest of the imaging data, each of the plurality of atlases having at least one image element with associated location and property information;

co-registering the plurality of atlases to the imaging data;

assigning a probability to generate a labeling parameter for the region of interest, the probability being associated with each atlas; and

classifying the region of interest of the subject based on said assigning.

1. A computer-implemented method for classifying a region of interest of a subject, comprising:

receiving imaging data comprising at least one image element, the imaging data comprising the region of interest of the subject;

providing a plurality of atlases, each of the plurality of atlases having a candidate region that corresponds to the region of interest of the imaging data, each of the plurality of atlases having at least one image element with associated location and property information;

co-registering the plurality of atlases to the imaging data, using at least one processor;

assigning a probability to generate a labeling parameter for the region of interest, the probability being associated with each atlas; and

classifying the region of interest of the subject based on said assigning.

29. The system of claim 28, wherein the probability for the labeling parameter is based on a plurality of probability terms of each atlas, the plurality of probability terms including a location value and a property value of the at least one image element of each atlas.

30. The system of claim 29, wherein the location value is selected from a most probable location assignment among the plurality of atlases and wherein the property value is selected from a most probable intensity assignment from the plurality of atlases.

31. The system of claim 29, wherein the classifying includes using a predetermined formula to combine the plurality of probability terms.

32. The system of claim 31, wherein:

the co-registering the plurality of atlases generates a mixture of probabilities from each atlas of at least one labeling parameter,

the generated mixture of probabilities is maximized with respect to the parameter being estimated, and

the mixture is generated via superposition and weighting of the probability from each atlas, and

the generated mixture is used in generating the labeling parameter.

33. The system of claim 32, wherein the plurality of probability terms is generated from statistical noise modeling of an observed imager modality of the imaging data, the statistical noise modeling determining the location values and the property values.

34. The system of claim 28, wherein the probability used in the classifying meets a predetermined statistical significance.

35. The system of claim 28, wherein for each of the plurality of atlases, the classifying incorporates:

a weight corresponding to a function of a modality of the imaging data and the atlas.

42. The system of claim 28, wherein for each of the plurality of atlases, the classifying incorporates:

a weight corresponding to a function of deformation of the image element from being mapped to the atlas.

46. The system of claim 28, wherein for each of the plurality of atlases, the classifying incorporates:

global anatomical features in the image such as a size of the brain and ventricles or intensities of the brain and non-brain tissues.

47. The system of claim 28, wherein at least one of the plurality of atlases are co-registered to the imaging data using a linear transformation, a nonlinear transformation, and/or a diffeomorphic transformation.

48. The system of claim 28, wherein the at least one image element is at least one of a pixel, a plurality of pixels, a voxel, or a plurality of voxels.

49. The system of claim 28, further comprising:

suggesting that the subject has an anatomical feature based on the classifying.

51. The system of claim 28, wherein the imaging data is generated from at least one of magnetic resonance imaging (MRI), computed tomography (CT), positron emission tomography (PET), ultrasound, or nuclear tracer three-dimensional imaging.

52. The system of claim 28, wherein the region of interest comprises a portion of at least one of a brain, a heart, a liver, skin, a lung, another organ, one or more bones, or any combination thereof.

53. The system of claim 28, wherein a difference in intensities between the candidate regions of interest of the plurality of atlases and the region of interest of the imaging data is determined by determining a distance vector.

33-1. The system of claim 28, wherein the labeling parameter is at least one of:

size of tissue, amount of cerebrospinal fluid or type of tissue.

36. The system of claim 36, wherein the weight increases as a difference in intensity between the region of interest and the corresponding candidate region of the atlas decreases.

37. The system of claim 36, wherein the weight determined using an expectation-maximization algorithm.

38. The system of claim 36, wherein the function of the modality is determined by an MRI target image modality, a measure of intensity, a vector of intensities, a matrix of intensities, or a finite alphabet valued label such as a semantic label arising from an ontology and atlas image intensity and/or a vector or a matrix of intensities and/or a finite alphabet valued label such as a semantic label arising from the ontology.

39. The system of claim 36, wherein the imaging data includes a plurality of images and the receiving includes receiving the plurality of images, and

wherein the co-registering includes co-registering the plurality of images to the plurality of atlases.

41. The system of claim 39, wherein the measure of intensity is at least one of:

a mean intensity of the region of interest;

a sum of intensities of the region of interest;

a highest intensity of the region of interest;

a lowest intensity of the region of interest;

a vector or a matrix of any of these values; and

a finite alphabet semantic labeling from an ontology.

40. The system of claim 40, wherein the weight corresponds to a function of at least two modalities of the imaging data and the plurality of atlases.

43. The system of claim 43, wherein the weight increases as an amount of deformation of the region of interest from being mapped to the atlas decreases.

44. The system of claim 43, wherein the weight is also a function of other features that are predetermined between the imaging data and the atlas.

45. The system of claim 45, wherein the other features include non-image information of the subject and atlases, the non-image information including a function of an age of the subject of the imaging data and an age corresponding to the plurality of atlases and/or a diagnosis of the subject of the imaging data and a diagnosis corresponding to the plurality of atlases.

50. The system of claim 50, wherein the anatomical feature is at least one of an abnormality, a disease, a condition, a diagnosis, or any combination thereof.

55. A non-transitory computer readable storage medium for classifying a region of interest of a subject, the computer readable storage medium comprising instructions that when executed enable a computing system to perform:

receiving imaging data comprising at least one image element, the imaging data comprising the region of interest of the subject;

providing a plurality of atlases, each of the plurality of atlases having a candidate region that corresponds to the region of interest of the imaging data, each of the plurality of atlases having at least one image element with associated location and property information;

co-registering the plurality of atlases to the imaging data, using at least one processor;

assigning a probability to generate a labeling parameter for the region of interest, the probability being associated with each atlas; and

classifying the region of interest of the subject based on said assigning.

56-81. -81. (canceled)