# US 10,872,415 B2 · Learning-based spine vertebra localization and segmentation in 3D CT

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 10,872,415 B2
- Document ID: US10872415B2
- Status: Granted
- Kind: B2
- Application number: 15/763,797
- Filed: 2016-10-25
- Published: 2018-10-04
- Granted: 2020-12-22
- Earliest claimed: 2015-10-29
- Assignees: Broncus Medical Inc.
- Applicants: Broncus Medical Inc.
- Inventors: Erkang Cheng; Lav Rai; Henky Wibowo
- Web page: https://patentgrep.com/patent/US10872415B2

## CPC

- G06T 7/11: Region-based segmentation
- G06T 7/143: involving probabilistic approaches, e.g. Markov random field [MRF] modelling
- G06T 7/162: involving graph-based methods
- G06T 2207/10081: Computed x-ray tomography [CT]
- G06T 2207/20076: Probabilistic image processing
- G06T 2207/20081: Training; Learning
- G06T 2207/30012: Spine; Backbone
- G06T 2207/30172: Centreline of tubular or elongated structure
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL

## Abstract

Described herein is a novel method and system for segmentation of the spine using 3D volumetric data. In embodiments, a method includes an extracting step, localization step, and segmentation step. The extracting step comprises detecting the spine centerline and the spine canal centerline. The localization step comprises localizing the vertebra and intervertebral disc centers. Background and foreground constraints are created for each vertebra digit. Segmentation is performed for each vertebra digit and based on the hard constraints.

## Claims

### Claim 1

A method for identifying an anatomic structure of a patient comprising:
receiving a 3D image data set of the patient;
detecting a spine centerline and a spine canal centerline of the anatomic structure from the 3D image data set of the patient wherein the detecting is based on a prediction map, and wherein the anatomic structure is the spine and comprises a plurality of interconnected spine digits;
localizing a vertebra center of each of the plurality of interconnected spine digits, wherein each said vertebra center is required to be on the spine centerline;
constructing hard constraints for each spine digit, wherein the hard constraints include digit-specific background and foreground constraints based on the spine canal centerline, vertebra centers, intervertebral disc centers, and the spine centerline; and
automatically segmenting the anatomic structure based on the hard constraints from the constructing step; and
displaying the segmented anatomical structure.

### Claim 2 (depends on 1)

The method of claim 1 comprising localizing the intervertebral disc center of each intervertebral disc, wherein each said intervertebral disc center is required to be on the spine centerline.

### Claim 3 (depends on 2)

The method of claim 2 wherein the background constraints for each spine digit comprise two cone-like regions.

### Claim 4 (depends on 1)

The method of claim 1 wherein the detecting is performed based on optimizing prediction values of the prediction map.

### Claim 5 (depends on 4)

The method of claim 4 wherein the optimizing prediction values of the prediction map is based on a shortest path algorithm.

### Claim 6 (depends on 1)

The method of claim 1 further comprising computing an augmented constrained region encompassing the anatomic structure, and based on the spine centerline and spine canal centerline arising from the detecting step.

### Claim 7 (depends on 1)

The method of claim 1 wherein the localizing step is estimated by a probabilistic inference algorithm.

### Claim 8 (depends on 1)

The method of claim 1 wherein the prediction map is based on a machine learning algorithm.

### Claim 9 (depends on 1)

The method of claim 1 wherein the anatomic structure is rigid.

### Claim 10 (depends on 1)

The method of claim 1 further comprising identifying the anatomical structure and wherein the identifying is performed by a computer and based on an annotated exemplary anatomical structure.

### Claim 11

A system for segmenting an anatomical structure of a patient, the system comprising:
a memory unit for storing 3D image data of the patient;
a programmed processor operable to:
detect a characteristic feature of the anatomical structure and compute an augmented constrained region encompassing the anatomical structure, and wherein the characteristic feature comprises at least one centerline selected from the group consisting of a spine centerline and a spine canal centerline and the detecting is based on a prediction map;
localize a vertebra center and an intervertebral disc center based on the augmented constrained region; and
segment the anatomical structure based on the detected characteristic feature and the augmented constrained region; and
a display in communication with the processor and for displaying the segmented anatomical structure.

### Claim 12 (depends on 11)

The system of claim 11 wherein the programmed processor is operable to construct hard constraints for a spine digit based on each vertebra center and disc center adjacent said spine digit, and to segment the anatomical structure based on the hard constraints.

### Claim 13 (depends on 12)

The system of claim 12, wherein the hard constraints comprise background constraints for each spine digit.

### Claim 14 (depends on 13)

The system of claim 13, wherein the background constraints comprise two cone-like regions.

### Claim 15 (depends on 11)

The system of claim 11 wherein computing the augmented constrained region computes a tube-like region encompassing the spine centerline and spine canal centerline.

### Claim 16 (depends on 11)

The system of claim 11 wherein the processor is further operable to identify the anatomical structure based on an annotated exemplary anatomical structure.

### Claim 17 (depends on 11)

The system of claim 11 wherein the processor is operable to detect the characteristic feature based on a prediction map, and the prediction map is based on a machine learning algorithm.