Image Processing Apparatus Region Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing medical image processing techniques face challenges in achieving high reproducibility and accuracy in extracting regions of interest, as they often rely on single reference points or inconsistent algorithm designs, leading to variable and low reproducibility in region extraction results.
Innovation Solution
An image processing apparatus that sets multiple reference points in an image, obtains initial extraction results based on these points, and integrates pixel values to extract the object region with high reproducibility and accuracy, using methods like level-set and graph-cut techniques to refine the extraction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a single reference point is used for region extraction, then the processing complexity is low, but the reproducibility and accuracy of region extraction are insufficient
Solution Approach 1:
The patent divides the region extraction process into multiple independent stages: (1) obtaining multiple candidate reference points, (2) generating multiple initial extraction results from each reference point, (3) integrating these results through pixel value addition, and (4) performing final extraction based on the integrated result. This segmentation allows each stage to contribute to overall reproducibility without requiring the entire system to be overly complex.
Solution Approach 2:
The patent merges multiple initial extraction results by adding pixel values from each result to create an integrated extraction result. This combining approach consolidates information from multiple reference points, improving reproducibility and accuracy while maintaining a systematic processing framework that manages complexity.
2Reliability
If multiple algorithms are integrated for region extraction, then the coverage of different extraction approaches is improved, but the reproducibility decreases due to inconsistent design policies
Solution Approach 1:
The patent creates a universal extraction framework that can accommodate multiple candidate reference points and generate multiple initial extraction results using a consistent design policy. This universal approach ensures that regardless of which reference points are selected or what characteristics are used, the same systematic process is applied, thereby maintaining reproducibility while allowing versatility in input selection.
Solution Approach 2:
The patent changes the parameter of reference point selection from a single point to multiple candidate points, and changes the extraction parameter from one algorithm execution to multiple executions with integration. This parameter transformation allows the system to maintain consistent design policies across multiple operations while improving reproducibility through aggregation.
3Measurement precision
If region extraction is performed without integrating multiple initial results, then the processing time is short, but the accuracy and reproducibility of the extracted region are low
Solution Approach 1:
The patent performs preliminary actions by obtaining multiple candidate reference points and generating multiple initial extraction results before the final extraction decision. This preliminary multi-result generation, followed by integration through pixel value addition, prepares comprehensive data that improves accuracy while allowing the final extraction step to efficiently utilize the integrated information.
Solution Approach 2:
The patent creates multiple copies of the extraction process, each starting from a different candidate reference point. These copied extraction results are then integrated by adding pixel values, allowing the system to leverage multiple copies to improve accuracy and reproducibility while the integration process efficiently consolidates the information from all copies.
Data Source
AI summary
An image processing apparatus is configured to extract an object region from an image. The image processing apparatus includes: a setting unit configured to set a plurality of reference points in the image; an obtaining unit configured to obtain a contour of the object region corresponding to each of the plurality of reference points as an initial extraction result based on a characteristic of the object region; and an extraction unit configured to extract the object region from the image based on an integration result obtained by integrating values of pixels in a plurality of initial extraction results.