Neural Network Medical Image Correction

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing correction algorithms for medical projection data are time-consuming, leading to inefficiencies in correcting artifacts and noise in medical images, which can impact diagnosis accuracy.

Innovation Solution

A method using a neural network model to determine correction coefficients and noise reduction parameters, trained with sample data and projection data, to rapidly correct artifacts and noise in medical images.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If existing correction algorithms are used to correct projection data, then the accuracy of artifact and noise correction is improved, but the processing time increases significantly

Engineering Contradiction:

Improvecorrection accuracyVSAvoidprocessing time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The patent replaces traditional iterative correction algorithms with a neural network model that has been pre-trained on simulation data. The neural network directly maps projection data to corrected data in a single forward pass, eliminating the need for time-consuming iterative computations while achieving comparable or superior correction accuracy.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The neural network model is pre-trained using simulation data generated from known ground truth projection data and corresponding corrected data. This preliminary training allows the model to learn optimal correction transformations in advance, enabling rapid inference on actual medical images without requiring time-consuming real-time computation.

Inventive Principle:

Principle #10Preliminary action

2Manufacturing precision

If traditional correction algorithms are applied to medical projection data, then correction precision is improved, but processing speed deteriorates

Engineering Contradiction:

Improvecorrection precisionVSAvoidprocessing speed

Core Design Contradiction:

Manufacturing precisionVSProductivity

Solution Approach 1:

The patent substitutes traditional iterative mathematical correction algorithms with a neural network-based system. The neural network performs correction operations through parallelizable matrix multiplications and activations, achieving both high precision and fast processing speeds that are incompatible with conventional sequential algorithms.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent transforms the correction process from parameter estimation through iterative optimization to direct parameter lookup and transformation using the neural network. The network outputs corrected projection data directly as a function of input data and pre-computed lookup tables, dramatically increasing processing speed while maintaining precision.

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS20240185486A1Systems and methods for determining parameters for medical image processing

Publication Date: 2024.06.06 SHANGHAI UNITED IMAGING HEALTHCARE

AI summary

A system and method for determine a parameter for medical data processing are provided. The method may include obtaining sample data, the sample data may comprise at least one of projection data or a scanning parameter. The method may also include obtaining a first neural network model. The method may further include determining the parameter based on the sample data and the first neural network model. The parameter may comprise at least one of a correction coefficient or a noise reduction parameter.