# US 2021/0279863 A1 · Data processing apparatus and method

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

- Number: US 2021/0279863 A1
- Document ID: US20210279863A1
- Status: Published application
- Kind: A1
- Application number: 16/992,466
- Filed: 2020-08-13
- Published: 2021-09-09
- Earliest claimed: 2020-03-03
- Assignees: The University Court of the University of Edinburgh; Canon Medical Systems Corporation
- Applicants: The University Court of the University of Edinburgh; Canon Medical Systems Corporation
- Inventors: Grzegorz JACENKÓW; Sotirios TSAFTARIS; Brian MOHR; Alison O'NEIL; Aneta LISOWSKA
- Web page: https://patentgrep.com/patent/US20210279863A1

## CPC

- G06T 5/70: Denoising; Smoothing
- G06T 7/0012: {Biomedical image inspection}
- G06T 5/20: using local operators
- G06T 5/60: using machine learning, e.g. neural networks
- G06T 2207/10072: Tomographic images
- G06T 2207/10116: X-ray image
- G06T 2207/20024: Filtering details
- G06T 2207/20081: Training; Learning
- G06T 2207/20084: Artificial neural networks [ANN]
- G06T 2207/20132: Image cropping
- G06T 2207/30004: Biomedical image processing
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL

## Abstract

A medical image data processing apparatus comprises processing circuitry configured to: receive medical image data in respect of at least one subject; receive non-image data; generate a filter based on the non-image data; and apply the filter to the medical image data, wherein the filter is configured to limit a region of the medical image data.

## Claims

### Claim 1

A medical image data processing apparatus comprising processing circuitry configured to:
receive medical image data in respect of at least one subject;
receive non-image data;
generate a filter based on the non-image data; and
apply the filter to the medical image data, wherein the filter is configured to limit a region of the medical image data.

### Claim 2 (depends on 1)

The apparatus according to claim 1, wherein the processing circuitry is configured to train a model,
the model comprising a plurality of layers,
the medical image data comprise sets of training data, and
the applying of the filter comprises applying the filter with respect to outputs of a layer of the model as part of the training of the model.

### Claim 3 (depends on 2)

The apparatus according to claim 2, wherein:
the model comprises a neural network;
the generating of the filter comprises training an auxiliary model to obtain at least one parameter value for the filter; and
the auxiliary model comprises an auxiliary neural network.

### Claim 4 (depends on 2)

The apparatus according to claim 2, wherein the applying of the filter to outputs of a layer of the model comprises applying the filter as part of a further layer of the model, the outputs of said layer being used as inputs to said further layer.

### Claim 5 (depends on 4)

The apparatus according to claim 4, wherein said further layer comprises a plurality of feature maps and/or channels, and the applying of the filter comprises using different parameter values and/or different functions for different ones of the feature maps and/or channels.

### Claim 6 (depends on 2)

The apparatus according to claim 2, wherein the method comprises determining respective parameter values of the filter for each of a plurality of channels and/or feature maps.

### Claim 7 (depends on 2)

The apparatus according to claim 2, wherein the processing circuitry is configured to at least one of:
a) determine parameter values for a plurality of filters, and applying the filters in respect of outputs of a plurality of different layers of the model as part of the training of the model; or
b) apply a process that encourages a variation in the position of peak or other characteristic of the filter for different ones of a plurality of channels and/or feature maps.

### Claim 8 (depends on 3)

The apparatus according to claim 3, wherein the determining of the parameter value(s) for the filter comprises learning the parameter value(s) separately for each of a plurality of channels and/or feature maps using the auxiliary neural network.

### Claim 9 (depends on 1)

The apparatus according to claim 1, wherein the non-image data comprises or represents at least one of age, weight, sex, presence or absence of a particular medical condition, at least one property of an imaging procedure or imaging apparatus, output(s) of another method, electronic health record(s) (EHR), text data, structured data, metadata, DICOM data or DICOM metadata.

### Claim 10 (depends on 1)

The apparatus according to claim 1, wherein the filter is a Gaussian filter.

### Claim 11 (depends on 10)

The apparatus according to claim 10, wherein the processing circuitry is further configured to determine parameters of the Gaussian filter based on the non-image data.

### Claim 12 (depends on 11)

The apparatus according to claim 11, wherein the processing circuitry is further configured to determine a peak position and variance of the Gaussian filter based on the non-image data.

### Claim 13 (depends on 11)

The apparatus according to claim 11, wherein the processing circuitry is further configured to provide a neural network which outputs parameters of the Gaussian filter based on the non-image data as input.

### Claim 14 (depends on 1)

The apparatus according to claim 1, wherein the generating of the filter comprises applying a process that encourages the filter to be non-spatially uniform.

### Claim 15 (depends on 14)

The apparatus according to claim 14, wherein the applying of a process that encourages the filter to be non-spatially uniform comprises using a penalty in a loss function.

### Claim 16 (depends on 1)

The apparatus according to claim 1, wherein the:
a) the filter represents a position of a feature of interest;
b) the filter represents relative positions of two or more features of interest.

### Claim 17 (depends on 1)

The apparatus according to claim 1, further comprising displaying a visual representation of the filter on an image obtained from an image data set.

### Claim 18 (depends on 1)

The apparatus according to claim 1, wherein at least one of:
a) the filter comprises a differentiable function and/or a compact or efficiently parameterized smooth function;
b) the filter comprises a Gaussian function or a Student's T-distribution;
c) the filter comprises a product of at least two vectors, optionally at least two Gaussian vectors, and/or a matrix;
d) the non-image data comprises at least one of image data, text data, metadata, audio data and/or structured data.

### Claim 19 (depends on 2)

The apparatus according to claim 2, comprising applying the trained model to a data set to produce an output.

### Claim 20

A method of processing medical image data comprising:
receiving medical image data in respect of at least one subject;
receiving non-image data;
generating a filter based on the non-image data; and
applying the filter to the medical image data, wherein the filter is configured to limit a region of the medical image data.
