# US 12,008,689 B2 · Devices, systems, and methods for deep-learning kernel-based scatter estimation and correction

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

- Number: US 12,008,689 B2
- Document ID: US12008689B2
- Status: Granted
- Kind: B2
- Application number: 17/542,245
- Filed: 2021-12-03
- Published: 2023-06-08
- Granted: 2024-06-11
- Earliest claimed: 2021-12-03
- Assignees: Canon Medical Systems Corporation
- Applicants: Canon Medical Systems Corporation
- Inventors: Yujie Lu; Tzu-Cheng Lee; Liang Cai; Jian Zhou; Zhou Yu
- Web page: https://patentgrep.com/patent/US12008689B2

## CPC

- G06T 12/10: Image preprocessing, e.g. calibration, positioning of sources or scatter correction
- G06N 3/045: Combinations of networks
- G06N 3/08: Learning methods
- G06N 3/084: Backpropagation, e.g. using gradient descent
- G16H 30/20: for handling medical images, e.g. DICOM, HL7 or PACS
- G16H 30/40: for processing medical images, e.g. editing
- G16H 50/20: for computer-aided diagnosis, e.g. based on medical expert systems
- G16H 50/50: for simulation or modelling of medical disorders
- G06T 2211/441: AI-based methods, deep learning or artificial neural networks
- G06T 2211/452: involving suppression of scattered radiation or scatter correction
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G16H: HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA

## Abstract

Devices, systems, and methods obtain first radiographic-image data reconstructed based on a set of projection data acquired in a radiographic scan; apply one or more trained machine-learning models to the set of projection data and the first radiographic-image data to obtain a set of parameters for a scatter kernel; input the set of parameters and the set of projection data into the scatter kernel to obtain scatter-distribution data; and perform scatter correction on the set of projection data using the scatter-distribution data, to obtain a set of corrected projection data.

## Claims

### Claim 1

A data processing method for scatter correction, the data processing method comprising:
obtaining first radiographic image data reconstructed based on a set of projection data acquired in a radiographic scan;
applying one or more trained machine-learning models to the set of projection data and the first radiographic image data to obtain a set of parameters for a scatter kernel;
inputting the set of parameters and the set of projection data into the scatter kernel to obtain scatter-distribution data; and
performing scatter correction on the set of projection data using the scatter-distribution data, to obtain a set of corrected projection data.

### Claim 2 (depends on 1)

The data processing method of claim 1, further comprising:
reconstructing a second radiographic image from the set of corrected projection data.

### Claim 3 (depends on 2)

The data processing method of claim 2, further comprising:
outputting the second radiographic image for display.

### Claim 4 (depends on 1)

The data processing method of claim 1, wherein the set of projection data are computed-tomography (CT) data,
wherein the set of corrected projection data are CT data, and
wherein the radiographic scan is an X-ray CT scan.

### Claim 5 (depends on 1)

The data processing method of claim 1, wherein the scatter kernel is a convolutional kernel.

### Claim 6 (depends on 1)

The data processing method of claim 1, wherein the one or more trained machine-learning models are neural networks.

### Claim 7 (depends on 1)

The data processing method of claim 1, wherein the set of parameters includes parameters that depend on scanned geometry, equivalent monochromatic energy, an anti-scatter grid, and one or more scanned materials.

### Claim 8

A method for training one or more machine-learning models, the method comprising:
obtaining a training dataset for a scatter model, wherein the training dataset includes projection data, radiographic-image data, and target data, wherein the target data define a scatter distribution, and wherein the scatter model accepts, as inputs, a set of parameters and projection data;
applying the projection data and the radiographic-image data to one or more machine-learning models to generate the set of parameters for the scatter model;
using the scatter model to generate, based on the set of parameters and on the projection data, output scatter-distribution data;
applying the output scatter-distribution data and the target data to a loss function to obtain a loss value; and
updating the one or more machine-learning models based on the loss value.

### Claim 9 (depends on 8)

The method of claim 8, wherein each of the one or more machine-learning models is a respective neural network.

### Claim 10 (depends on 8)

The method of claim 8, wherein the target data define a scatter distribution.

### Claim 11 (depends on 8)

The method of claim 8, wherein obtaining the training dataset includes:
obtaining raw projection data and raw radiographic-image data;
performing first preprocessing on the raw radiographic-image data to generate the radiographic-image data; and
performing second preprocessing on the raw projection data to generate the projection data.

### Claim 12 (depends on 8)

The method of claim 8, wherein the one or more machine-learning models include at least two machine-learning models, and
wherein each machine-learning model of the at least two machine-learning models outputs one or more respective parameters of the set of parameters for the scatter model.

### Claim 13 (depends on 8)

The method of claim 8, wherein the loss value is a gradient of the loss function.

### Claim 14 (depends on 8)

The method of claim 8, wherein the scatter model is a convolutional kernel.

### Claim 15

An information processing apparatus comprising:
one or more computer-readable storage media; and
one or more processors, wherein the one or more processors cooperate with the one or more computer-readable storage media to cause the information processing apparatus to
obtain first radiographic-image data reconstructed based on a set of projection data acquired in a radiographic scan;
apply one or more trained machine-learning models to the set of projection data and the first radiographic-image data to obtain a set of parameters for a scatter model;
apply the scatter model, using the set of parameters and the set of projection data as inputs, to obtain scatter-distribution data; and
perform scatter correction on the set of projection data using the scatter-distribution data, to obtain a set of corrected projection data.

### Claim 16 (depends on 15)

The information processing apparatus of claim 15, wherein the one or more trained machine-learning models are neural networks.

### Claim 17 (depends on 15)

The information processing apparatus of claim 15, wherein the set of parameters includes parameters that depend on scanned geometry, equivalent monochromatic energy, an anti-scatter grid, and one or more scanned materials.

### Claim 18 (depends on 15)

The information processing apparatus of claim 15, wherein the one or more processors cooperate with the one or more computer-readable storage media to further cause the information processing apparatus to
reconstruct a second radiographic image from the set of corrected projection data.

### Claim 19 (depends on 15)

The information processing apparatus of claim 15,
wherein the one or more trained machine-learning models include a first trained machine-learning model that accepts both projection data and radiographic-image data as inputs.

### Claim 20 (depends on 15)

The information processing apparatus of claim 15,
wherein the one or more trained machine-learning models include a first trained machine-learning model that accepts projection data, but not radiographic-image data, as inputs; and
wherein the one or more trained machine-learning models include a second trained machine-learning model that accepts radiographic-image data, but not projection data, as inputs.