Automatic 3D Medical Image Orientation via Anchor-Organ Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining the orientation of medical images are inefficient, require feature engineering, depend on modality or body part size, lack invariance to affine transforms, and are not compatible with modern image segmentation datasets, making it difficult to align and normalize medical images for clinical use.
Innovation Solution
A method involving the selection of anchor organs, training a segmentation model to identify these organs, generating a segmentation mask, computing image coordinates, and aligning images based on a correlation between image and anatomical coordinates using a decision function for coordinate system transition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual rotation and alignment of medical images is performed by clinical experts, then image orientation can be corrected, but time consumption increases significantly
Solution Approach 1:
The system performs automatic image alignment by having the medical images self-correct their orientation through algorithmic processing. The coordinate system transformation and anchor organ-based alignment enable the images to determine their own orientation without manual intervention, eliminating the time-consuming manual rotation process while maintaining accurate orientation correction
Solution Approach 2:
The manual mechanical rotation operation is replaced with an automated computational system. The patent substitutes the physical manual manipulation of images with a digital coordinate transformation system that uses anchor organs and decision functions to automatically compute and apply the necessary rotations and translations, achieving both speed and precision
2Loss of information
If existing landmark localization approaches are used, then orientation information can be extracted, but the methods require feature engineering and are not compatible with modern segmentation datasets
Solution Approach 1:
The patent creates a universal alignment system based on anchor organs that works across different imaging modalities and dataset types. By using fundamental anatomical landmarks (anchor organs) that exist in all medical images, the system achieves compatibility with modern segmentation datasets without requiring modality-specific feature engineering, making the solution universally applicable
Solution Approach 2:
The patent extracts only the essential orientation-determining elements (anchor organs) from the complex medical images. By isolating and using only these key anatomical landmarks for coordinate system transformation, the system eliminates the need for comprehensive feature engineering while still capturing all necessary orientation information
3Measurement precision
If existing landmark localization methods are applied, then body orientation can be determined, but computational efficiency decreases due to complex processing requirements
Solution Approach 1:
The patent segments the image processing task by identifying and using only specific anchor organs for orientation determination. This segmentation approach divides the complex full-image analysis into focused processing of key anatomical landmarks, reducing computational load while maintaining precise body orientation determination through the use of these strategically selected reference points
Data Source
AI summary
A system, method and non-transitory computer readable storage medium for aligning a set of medical images. The operations for aligning the medical images include receiving a set of medical images, selecting anchor organs from the medical images, training a segmentation model to identify the anchor organs in the medical images based on a training dataset and generating, based on the segmentation model, a segmentation mask for the anchor organs. The operations also include computing image coordinates for the anchor organs in each of the medical images based on a center of mass of each of the anchor organs, determining a correlation between the image coordinates for the anchor organs in each of the medical images and corresponding anatomical coordinates for the anchor organs in the training dataset and aligning each of the set of medical images based on the correlation between the image coordinates and the anatomical coordinates.