X-ray Image Registration Error Labeling via Sensor Feedback

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing image stitching technologies in X-ray imaging systems rely heavily on personal experience, leading to inaccuracies and inefficiencies in determining stitching errors, which affects the accuracy and time required for image registration.

Innovation Solution

An image processing method and apparatus that calculates relative displacement vectors between images, uses position sensor feedback to determine error, and applies a pre-stored training model to assess registration accuracy, enabling automated labeling and improving stitching precision.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Area of stationary object

If image stitching technology is used to expand the imaging field of view, then the field of view is improved, but the accuracy of image registration deteriorates due to reliance on personal experience

Engineering Contradiction:

Improvefield of viewVSAvoidaccuracy of image registration

Core Design Contradiction:

Area of stationary objectVSMeasurement precision

Solution Approach 1:

The patent implements feedback by calculating the difference between the first displacement vector (from image registration) and the second displacement vector (from position sensor) to obtain a first error. This error is then used to evaluate the registration accuracy and provide feedback for improving the stitching process, replacing reliance on personal experience with objective measurement feedback.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The patent replaces the mechanical/manual method of judging stitching errors by doctor's personal experience with an automated computational system. The system uses displacement vector calculations, position sensor feedback, and machine learning models to automatically evaluate registration accuracy, substituting human judgment with an automated measurement and evaluation system.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Ease of operation

If manual judgment by doctor's personal experience is used to determine stitching errors, then the method is simple to implement, but the productivity deteriorates due to greater time and effort required

Engineering Contradiction:

Improvesimplicity of implementationVSAvoidworking efficiency

Core Design Contradiction:

Ease of operationVSProductivity

Solution Approach 1:

The patent implements self-service by enabling the system to automatically evaluate its own registration accuracy without requiring manual intervention. The system calculates displacement vectors, compares them with position sensor feedback, uses machine learning models to assess errors, and automatically labels registration levels, making the evaluation process self-contained and eliminating the need for doctor's personal judgment.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent changes the evaluation parameters from subjective personal experience to objective quantitative measurements. By introducing displacement vectors, position sensor readings, error calculations, and machine learning model outputs, the system transforms the evaluation from a qualitative manual process to a quantitative automated process, improving both accuracy and efficiency.

Inventive Principle:

Principle #35Parameter changes

3Productivity

If automated image registration is used to improve working efficiency, then the productivity is improved, but the measurement precision deteriorates without proper error evaluation mechanisms

Engineering Contradiction:

Improveworking efficiencyVSAvoidaccuracy of registration evaluation

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The patent applies preliminary action by pre-storing training models that contain mathematical distribution models of second errors between multiple third displacement vectors and fourth displacement vectors. These pre-trained models enable the system to accurately evaluate registration accuracy without requiring manual intervention, ensuring both automation and precision.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS10089747B2Image processing method and apparatus for X-ray imaging device

Publication Date: 2018.10.02 GE PRECISION HEALTHCARE LLC

AI summary

This disclosure presents an image processing method and related X-ray imaging device The method comprises: calculating a relative displacement between two first images that are already in auto registration as a first displacement vector; calculating a difference between position information fed back by a position sensor on the X-ray imaging device when imaging exposure is performed on the two first images respectively as a second displacement vector; calculating a first error of the first displacement vector relative to the second displacement vector; calculating a registration level corresponding to the first error in accordance with a pre-stored training model which is a mathematical distribution model of second errors between a plurality of third displacement vectors and a plurality of corresponding fourth displacement vectors; and labeling the registration level on the two first images that are already in auto registration.