# US 2019/0377978 A1 · Knowledge-based ultrasound image enhancement

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

- Number: US 2019/0377978 A1
- Document ID: US20190377978A1
- Status: Published application
- Kind: A1
- Application number: 16/546,423
- Filed: 2019-08-21
- Published: 2019-12-12
- Earliest claimed: 2015-05-27
- Assignees: Siemens Medical Solutions USA, Inc.
- Applicants: Siemens Medical Solutions USA, Inc.
- Inventors: Bimba Rao; Helene Houle; Bogdan Georgescu
- Web page: https://patentgrep.com/patent/US20190377978A1

## CPC

- A61B 8/0891: {for diagnosis of blood vessels}
- A61B 8/13: Tomography (A61B8/10, A61B8/12 take precedence)
- A61B 8/463: {characterised by displaying multiple images or images and diagnostic data on one display}
- A61B 8/5269: {involving detection or reduction of artifacts}
- G06F 18/214: Generating training patterns; Bootstrap methods, e.g. bagging or boosting
- G06T 5/60: using machine learning, e.g. neural networks
- G06V 10/774: Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
- G06T 2207/10132: Ultrasound image
- G06T 2207/20012: Locally adaptive
- G06T 2207/20076: Probabilistic image processing
- G06T 2207/20081: Training; Learning
- G06T 2207/20192: Edge enhancement; Edge preservation
- G06T 2207/30004: Biomedical image processing
- A61B: DIAGNOSIS; SURGERY; IDENTIFICATION (usefulness limited to only animals A61D)
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING

## Abstract

Image enhancement is provided for medical diagnostic ultrasound. Knowledge-based detection of anatomy or artifact identifies locations to be enhanced. The knowledge-based detection of the locations may avoid identification of other anatomy or artifacts. The image enhancement is applied to the identified locations and not others.

## Claims

### Claim 1

A method of image enhancement in medical diagnostic ultrasound, the method comprising:
acquiring, by an ultrasound system, ultrasound image data from a scan of tissue of a patient, the ultrasound image data representing spatial locations of the tissue;
applying, by a processor of the ultrasound system, a machine-learnt classifier to the ultrasound image data, the machine-learnt classifier having been trained with machine learning from a database of ground-truth annotated images, the machine-learnt classifier outputting locations of an imaging artifact of the ultrasound system in the ultrasound image data in response to input of the ultrasound image data to the machine-learnt classifier;
altering the ultrasound image data for the locations of the imaging artifact differently than for other locations; and
displaying an ultrasound image of the tissue of the patient from the altered ultrasound image data.

### Claim 2 (depends on 1)

The method of claim 1 wherein acquiring comprises acquiring B-mode image data after B-mode detection.

### Claim 3 (depends on 1)

The method of claim 1 wherein applying comprises:
extracting input feature values from the ultrasound image data; and
outputting the locations in response to input of the input feature values to the machine-learnt classifier.

### Claim 4 (depends on 1)

The method of claim 1 wherein applying comprises detecting the locations of the imaging artifact as locations of a grating lobe, and wherein altering comprises applying a filter or amplitude scaling to the location of the imaging artifact and not at the other locations.

### Claim 5 (depends on 1)

The method of claim 1 wherein applying comprises detecting the locations of the imaging artifact as locations of spontaneous contrast of blood cells, and wherein altering comprises amplitude scaling more greatly for or applying greater temporal persistence to the locations of the imaging artifact than the other locations.

### Claim 6 (depends on 1)

The method of claim 1 wherein applying comprises detecting shadowing, and wherein altering comprises reducing the shadowing with an increase in intensity in the shadowing.

### Claim 7 (depends on 1)

The method of claim 1 wherein altering comprises spatially adaptive filtering, the spatially adaptive filtering adapting as a function of the locations of the imaging artifact.

### Claim 8 (depends on 1)

The method of claim 1 further comprising:
outputting an indication of the locations of the imaging artifact; and
performing the altering after receiving user confirmation based on the output indication.

### Claim 9 (depends on 1)

The method of claim 1 further comprising:
detecting an anatomic structure represented in the ultrasound image data with an additional machine-learnt classifier; and
filtering the ultrasound image data based on the anatomic structure,
wherein displaying comprises displaying the ultrasound image from the altered and filtered ultrasound image data.

### Claim 10 (depends on 1)

The method of claim 1 wherein displaying comprises displaying the ultrasound image with the imaging artifact reduced relative to the ultrasound image without the altering.

### Claim 11

A non-transitory computer readable storage medium having stored therein data representing instructions executable by a programmed processor for image enhancement in medical diagnostic ultrasound, the storage medium comprising instructions for:
receiving, from an ultrasound scanner, detected ultrasound data representing a patient;
classifying locations represented by the detected ultrasound data as representing an artifact or anatomy, the classifying being with a knowledge base;
altering the detected ultrasound data as a function of the classification of the locations, the altering being an altering of amplitude of the ultrasound data for the artifact or anatomy differently relative to the amplitude of the ultrasound data for representing background, tissue, fluid and/or another object different than the classified artifact or anatomy while maintaining representation of all of the locations including representation of the background, tissue, fluid and/or the other object; and
generating an image from the enhanced ultrasound data as altered.

### Claim 12 (depends on 11)

The non-transitory computer readable storage medium of claim 11 wherein classifying comprises classifying with the knowledge base comprising ground-truth labeled database of ultrasound data.

### Claim 13 (depends on 11)

The non-transitory computer readable storage medium of claim 11 wherein classifying comprises classifying with a machine-learnt classifier, the machine-learnt classifier learnt from the database.

### Claim 14 (depends on 11)

The non-transitory computer readable storage medium of claim 11 wherein classifying comprises detecting an anatomical structure as the anatomy, and wherein altering comprises filtering as a function of the locations of the anatomical structure.

### Claim 15 (depends on 11)

The non-transitory computer readable storage medium of claim 11 wherein classifying comprises detecting the artifact, and wherein altering comprises reducing the amplitude for the locations of the artifact.

### Claim 16 (depends on 11)

The non-transitory computer readable storage medium of claim 11 wherein receiving comprises receiving B-mode data, wherein altering comprises filtering, and wherein generating the image comprises generating a B-mode image without segmentation.

### Claim 17

A system for image enhancement in medical diagnostic ultrasound, the system comprising:
a receive beamformer configured to acquire ultrasound data representing a region of a patient;
a B-mode detector, Doppler estimator, or both configured to output detected data from the ultrasound data;
a processor configured to extract input features from the detected data, identify an artifact from the detected data as a function of a knowledge base, and image process the detected data as a function of the artifact, the image process including processing the detected data for locations of the artifact differently than other locations; and
a display configured to display an image of the region based on the image processed detected data, the image being an ultrasound image with the artifact reduced.

### Claim 18 (depends on 17)

The system of claim 17 wherein the processor is configured to identify with a machine-learnt classifier derived from the knowledge base.
