Medical Image Processing Apparatus Error Correction via Blurred Region Masking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In medical image processing, two-step extraction methods can fail if errors occur in the first stage of region inference due to loss of region or boundary misrecognition, affecting the accuracy of target region extraction in the second stage.
Innovation Solution
A medical image processing apparatus that acquires an input image, infers a first region, generates a corrected image by blurring the inferred region to mitigate inference errors, and performs a second inference based on the corrected image to improve the accuracy of the target region extraction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If two-step extraction method is used to reduce search range and simplify learning processing, then productivity and device complexity are improved, but reliability deteriorates due to error propagation from first stage to second stage
Solution Approach 1:
The patent applies feedback by using the inferred first region image as an input to the second inference process. The second inference takes both the original medical image and the first region inference result as inputs, allowing the system to refine and correct errors from the first stage while maintaining the benefits of region narrowing. This feedback loop ensures that errors in boundary recognition or region loss in the first stage can be corrected in the second stage, thus improving reliability without sacrificing the productivity gains from the two-step approach.
2Device complexity
If first stage inference is performed to narrow down region before second stage extraction, then device complexity is reduced, but measurement precision deteriorates due to boundary misrecognition and region loss
Solution Approach 1:
The patent segments the extraction process into two distinct stages: first inference for region narrowing and second inference for precise target extraction. Each stage has a specialized function - the first stage focuses on reducing the search space by identifying candidate regions, while the second stage focuses on precise boundary detection and accurate target region extraction. This segmentation allows each stage to be optimized for its specific purpose, maintaining low complexity while improving precision through the specialized second stage.
Solution Approach 2:
The first inference stage performs preliminary action by pre-identifying candidate regions and narrowing down the search space before the second stage begins. This preliminary region masking reduces the computational burden on the second inference model, allowing it to focus computational resources on achieving high precision boundary detection and accurate target extraction within the pre-defined candidate regions, thus maintaining low overall complexity while achieving high precision.
Data Source
AI summary
An image processing apparatus according to an embodiment includes processing circuitry. The processing circuitry acquires an input image, infers a first region image about a first region included in the input image, generates a corrected image with the first region corrected from the first region image, and performs inference about a second region included in the input image based on the input image and the corrected image.