Trained Image Processing Model for 3D Medical Imaging Rendering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Physically based volume rendering (PBVR) techniques for medical imaging are time-consuming due to the need for a large number of ray tracing operations, making them inefficient for generating high-quality images quickly.
Innovation Solution
A system that uses a trained image processing model to generate a target rendered image by inputting an initial rendered image and characteristic parameters obtained through ray casting operations on 3D imaging data, improving image quality and processing speed by reducing the number of required ray tracing operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If a large number of ray tracing operations are performed to generate high-quality rendered images, then image quality is improved, but processing time increases
Solution Approach 1:
The patent performs ray casting operations in advance to pre-calculate characteristic parameters (such as color, opacity, depth) of the 3D imaging data. These pre-calculated parameters are then used by the trained image processing model to generate high-quality rendered images without requiring extensive real-time ray tracing operations, thus reducing processing time while maintaining image quality
Solution Approach 2:
The patent introduces a trained image processing model as an intermediary between the raw 3D imaging data and the final rendered image. This model learns the mapping from characteristic parameters to high-quality rendered images through training, enabling it to generate realistic rendered images quickly without performing numerous ray tracing operations during actual use
2Loss of information
If ray casting operations are performed to extract characteristic parameters, then image features are improved, but computational complexity increases
Solution Approach 1:
The patent extracts only the essential characteristic parameters (color, opacity, depth, normal vectors) from the 3D imaging data through ray casting operations, rather than performing complete ray tracing. This selective extraction obtains sufficient image features for realistic rendering while significantly reducing computational complexity
Data Source
AI summary
The present disclosure provides a system and method for image processing. The method may include obtaining an initial rendered image of a subject generated based on three-dimensional (3D) imaging data; obtaining one or more characteristic parameters corresponding to the 3D imaging data determined by performing a ray casting operation on the 3D imaging data; and generating a target rendered image of the subject by inputting the initial rendered image and the one or more characteristic parameters into a trained image processing model.