Dynamic Image Processing for Depth-Aware Lung Anatomy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image processing systems fail to superimpose anatomical information, such as lung lobes and depth, on each frame of a dynamic image, making it difficult to accurately predict the difficulty of lung resection surgery.
Innovation Solution
An image processing apparatus that integrates dynamic images with DRR images, adding depth information from anatomical data, allowing for the superimposition of anatomical information like lung lobes on dynamic images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If conventional image processing systems superimpose DRR images on dynamic images, then basic anatomical information can be displayed, but depth information and detailed anatomical structures (lung lobes, bronchus, blood vessels) cannot be visualized on each frame
Solution Approach 1:
The patent applies dimensionality change by extracting three-dimensional anatomical information from volume data and projecting it onto two-dimensional dynamic image frames. The extraction unit obtains 3D anatomical structures (lung lobes, bronchus, blood vessels) from CT volume data, and the superimposition unit projects these 3D structures onto 2D dynamic image frames with depth encoding, enabling visualization of depth information that would otherwise be lost in conventional 2D superimposition.
Solution Approach 2:
The patent segments the image processing system into distinct functional units: an extraction unit that separates anatomical information from volume data, a depth information generation unit that calculates depth metrics, and a superimposition unit that combines elements. This segmentation allows each unit to specialize in specific tasks, improving the quality of anatomical information extraction and depth encoding without overwhelming system complexity.
2Loss of information
If dynamic imaging is used to capture lung movement and breathing, then functional information is obtained, but anatomical information of lung lobes and structures cannot be sufficiently captured
Solution Approach 1:
The patent merges dynamic imaging data with static anatomical information from CT volume data. The superimposition unit combines the temporal-functional information from dynamic frames with the spatial-anatomical information from CT extraction, creating an integrated image that displays both lung movement dynamics and anatomical structures (lobes, bronchus, blood vessels) simultaneously, eliminating the need for separate imaging modalities.
Solution Approach 2:
The patent introduces DRR images as an intermediary that bridges dynamic imaging and CT anatomy. The DRR images serve as a mediator that can be superimposed on dynamic frames, carrying anatomical information from CT data while maintaining compatibility with dynamic image formats, thus enabling integration without direct complex interaction between CT and dynamic imaging systems.
3Loss of information
If multiple imaging modalities (CT, MRI, dynamic imaging) are used to capture complete anatomical and functional information, then comprehensive data is obtained, but individual checking of each image type is required
Solution Approach 1:
The patent merges multiple imaging modalities into a single integrated display. The superimposition unit combines dynamic image frames with DRR images containing anatomical information from CT data, creating composite images that simultaneously present functional dynamics and anatomical structures. This allows clinicians to review all necessary information (lung movement, lobes, bronchus, blood vessels) in one unified view rather than switching between separate imaging modalities.
Data Source
AI summary
Disclosed is an image processing apparatus including: a hardware processor that: acquires a dynamic image including a plurality of frame images by performing dynamic imaging on a target site of a subject by a radiation imaging apparatus; acquires a functional information image including volume data including a plurality of voxels by imaging the target site by another different modality; acquires anatomical information or lesion information regarding the target site of the functional information image; generates a DRR image based on the functional information image and the anatomical information or the lesion information; integrates a plurality of frames of the dynamic image and the DRR image to obtain an integrated image; and outputs the integrated image. The hardware processor adds, to the DRR image, depth information indicating a distance from a reference position of the anatomical information or the lesion information in the volume data.