Automated FISH Dot Counting via Multi-Level H-Maxima Transforms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Automatic FISH dot detection and counting in biomedical research is challenging due to background artifacts, dense clusters of cell nuclei, and low contrast, making existing algorithms inefficient and prone to errors.
Innovation Solution
A computer-implemented method for cell-level FISH dot counting using multi-level extended h-maxima or h-minima transforms, combined with wavelet-based segmentation and top-hat filtering, to extract FISH binary masks and accurately count dots within nuclei, while selecting optimal sensitivity levels for each cell based on contrast scores.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual counting of FISH dots is performed, then accuracy can be maintained through expert judgment, but the process becomes time-consuming and subjective
Solution Approach 1:
The patent replaces the manual mechanical counting process with an automated image processing system that uses mathematical morphology operations (top-hat filtering, h-maxima/h-minima transforms) and template matching algorithms to detect and count FISH dots, thereby eliminating time-consuming manual intervention while maintaining objective accuracy
Solution Approach 2:
The patent creates a digital copy of the FISH image and processes it through multiple computational steps including background subtraction, noise filtering, and pattern recognition algorithms to generate automated counting results, replacing the need for direct manual observation and counting
2Productivity
If existing automatic counting algorithms are used, then processing time is reduced, but accuracy deteriorates due to background artifacts, dense clusters, and low contrast
Solution Approach 1:
The patent segments the FISH image processing into distinct sequential steps: background subtraction using top-hat filtering, noise reduction through morphological operations, contrast enhancement via h-maxima/h-minima transforms at multiple levels, and final dot detection using template matching. This segmentation allows each step to optimize for its specific function, improving overall accuracy while maintaining automated processing speed
Solution Approach 2:
The patent extends the processing from single-level analysis to multi-level h-maxima/h-minima transforms, analyzing the image at different threshold levels and scales. This dimensional extension allows the algorithm to detect dots across varying contrast conditions and separate true signals from background artifacts more effectively
3Measurement precision
If multi-level extended h-maxima or h-minima transforms are applied to extract FISH binary masks, then detection accuracy in low-contrast images improves, but computational complexity increases
Solution Approach 1:
The patent applies preliminary background subtraction using top-hat filtering and noise reduction through morphological operations before applying the multi-level h-maxima/h-minima transforms. This preliminary processing simplifies the image data, reducing the computational burden of subsequent multi-level analysis while preserving the accuracy benefits of the extended transforms
Data Source
Figure 1A~2
Figure 3
Figure 4
AI summary
The invention relates to a computer implemented method and systems for cell level fish dot counting. FISH (fluorescence in situ hybridization) dot counting is the process of enumerating chromosomal abnormalities in the cells which can be used in areas of diagnosis and cancer research. The method comprises in part overlaying images of a biological sample comprising a nuclear counterstain mask and a FISH binary mask. The FISH binary mask is extracted using a multi-level extended h-maxima or h-minima.