CBCT Head Motion Correction With Prior 3D Bone Segmentation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current cone-beam CT systems are prone to patient motion artifacts due to slower projection data acquisition, leading to distorted reconstructions with streaks, blurring, and double structures, and existing correction methods either fail to converge or take too long, hampering workflow efficiency.

Innovation Solution

An image processing system that utilizes a reconstructed reference image with reduced motion artifacts to correct target images by aligning acquired projections with computed forward projections, leveraging bone segmentation and spatial registration to estimate and correct motion, allowing for faster and more robust convergence.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If iterative reconstruction approaches are used for motion correction, then motion artifacts can be reduced, but convergence is slow and unreliable

Engineering Contradiction:
Improvemotion artifact correction reliabilityVSAvoidconvergence time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent applies preliminary action by performing bone segmentation on a reference image before the actual CBCT reconstruction. This pre-processed bone map is then used to guide the reconstruction algorithm, providing motion correction information in advance and enabling faster, more reliable convergence without requiring extensive iterative processing during the main reconstruction phase.

Inventive Principle:
Principle #10Preliminary action

2Productivity

If conventional CBCT reconstruction is used, then imaging speed is maintained, but motion artifacts severely degrade image quality

Engineering Contradiction:
Improveimaging speedVSAvoidimage quality
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The patent applies segmentation by dividing the imaging problem into distinct components: bone structures and soft tissues. By segmenting the bone from the reference image and using it as a separate motion reference, the system can correct for head motion independently while maintaining fast CBCT reconstruction speeds for the complete image.

Inventive Principle:
Principle #1Segmentation

3Loss of time

If a prior reference image is utilized for motion correction, then convergence speed improves, but additional processing steps are required

Engineering Contradiction:
Improveconvergence timeVSAvoidprocessing workflow complexity
Core Design Contradiction:
Loss of timeVSDevice complexity

Solution Approach 1:

The patent uses an intermediary approach by introducing a bone map derived from the reference image as a mediator between the reference image and the CBCT reconstruction. This bone map serves as a simplified intermediate representation that captures motion information without requiring direct complex processing of the full reference image, thus reducing overall system complexity.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP4625328A1Head motion correction for cone-beam computed tomography (CBCT) reconstruction using bone segmentation from prior 3D imaging
Publication Date: 2025.10.01 KONINKLIJKE PHILIPS NV
  • EP4625328A1 patent drawingFigure 1
  • EP4625328A1 patent drawingFigure 2
  • EP4625328A1 patent drawingFigure 3

AI summary

An image processing system and related method. The image processing system (ARS) may comprise an input interface (IN) for receiving a reconstructed current target image (V) of an object of interest (OI) subjectable to motion. A motion corrector (MC) corrects for motion artifact in the target image causable by said motion, based on information as per a reconstructed reference image (V0) of the object. The said reference image (V0) may include motion artifact, if any, to a lesser extent than the current target image (V). The system allows for improved, that is faster and/or more robust, reduction of motion artifacts in the current target image.