Apparatus, Method and Computer Program
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing segmentation techniques for multiplexed biomarker images face challenges with intricate structures, noisy backgrounds, and quality issues, leading to inefficiencies and errors in analyzing spatial relationships between cells.
Innovation Solution
A segmentation-free process using distance transformation with multiple thresholds to generate structure image data, enhancing or suppressing pixel values based on proximity, thereby improving visibility and accuracy without the need for image segmentation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If segmentation techniques are used to analyze spatial relationships in multiplexed biomarker images, then cell interactions and associations can be identified, but the process encounters challenges with intricate structures, noisy backgrounds, and quality issues
Solution Approach 1:
The patent extracts the essential spatial relationship information by applying distance transformation to reference sample data, separating the structural framework from the complex image details. This allows spatial analysis to proceed without being hindered by noisy backgrounds and intricate structures that complicate segmentation
Solution Approach 2:
The patent transforms the reference sample data through distance transformation, changing the parameter representation from original intensity values to distance metrics. This parameter transformation enables robust spatial relationship analysis that is insensitive to background noise and image quality variations
2Measurement precision
If multiple thresholds are used for distance transformation, then the accuracy and robustness of structure image data generation is improved, but the computational complexity increases
Solution Approach 1:
The patent applies multiple thresholds for distance transformation, using more computational resources than a single threshold would require. This excessive action improves the accuracy and robustness of structure image data generation by capturing spatial relationships at multiple distance scales, ensuring reliable results even for complex biological structures
3Reliability
If segmentation-free process is used to enhance or suppress pixel values, then the visibility of sample structure is improved without errors from segmentation, but the method requires new computational approaches
Solution Approach 1:
The patent introduces structure image data as an intermediary representation that captures spatial relationships without requiring segmentation of the original images. This intermediary structure enables direct enhancement or suppression of pixel values based on spatial proximity, eliminating segmentation errors while providing a clear computational framework
Data Source
AI summary
Examples relate to an apparatus comprising one or more processors and one or more storage devices. The apparatus is configured to obtain reference sample data indicative of a first image of a sample from a first channel and to obtain threshold data indicative of a plurality of thresholds for performing a distance transformation. Further, the apparatus is configured to generate, using at least two thresholds of the plurality of thresholds, reference data by applying the distance transformation to the reference sample data. The apparatus is further configured to obtain sample data indicative of a second image of the sample from a second channel different from the first channel and to generate structure image data indicative of a structure of the sample. The structure image data is generated based on the reference data and the sample data.