Whole Slide Cell Counting Using Patch-Based Watershed Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current whole slide imaging systems struggle to efficiently count large numbers of cells, particularly one million or more cells, in a selected area of a pathology image, which is necessary for modern cancer diagnostics like NantOmics® Genomic Proteomic and Spectrometry (GPS) cancer diagnostics, and existing local computer-based methods are inadequate for real-time processing.
Innovation Solution
A client-server architecture utilizing whole slide stain separation, local drain cell segmentation, and watershed transform methods for patch-based cell counting, employing global and local thresholding filters, chamfer distance transforms, and watershed transformations to achieve real-time cell counting within a whole slide pathology image.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If local computer-based cell counting programs are used, then cell counting can be performed with simple equipment, but the processing speed is insufficient for counting one million or more cells in real-time
Solution Approach 1:
The patent divides the whole slide image into multiple patches and processes them in parallel. The image is segmented into regions of interest, and cell counting is performed on each patch independently using watershed transformation and distance map analysis, then results are aggregated to achieve real-time counting of one million or more cells
Solution Approach 2:
The patent transforms the 2D image data into a 3D distance map by calculating distance transforms. This dimensional transformation enables more accurate cell segmentation and counting by adding depth information to the flat image data, improving counting accuracy for large cell populations
2Measurement precision
If conventional thresholding methods are used for cell segmentation, then the process is simple and fast, but accuracy deteriorates when cells are densely packed or have varying intensities
Solution Approach 1:
The patent introduces a distance map as an intermediary representation between the original image and the final cell segmentation. The distance transform creates a topographic map where local maxima correspond to cell centers, serving as a mediator that simplifies the segmentation process while improving accuracy for densely packed cells
Solution Approach 2:
The patent replaces conventional intensity-based thresholding with a geometric approach using distance transforms and watershed algorithms. This substitution of the segmentation mechanism from intensity-based to geometry-based methods improves accuracy for cells with varying intensities and dense packing
Data Source
AI summary
Techniques are provided for determining a cell count within a whole slide pathology image. The image is segmented using a global threshold value to define a tissue area. A plurality of patches comprising the tissue area are selected. Stain intensity vectors are determined within the plurality of patches to generate a stain intensity image. The stain intensity image is iteratively segmented to generate a cell mask using a local threshold value that is and gradually reduced after each iteration. A chamfer distance transform is applied to the cell mask to generate a distance map. Cell seeds are determined on the distance map. Cell segments are determined using a watershed transformation, and a whole cell count is calculated for the plurality of patches based on the cell segments. A client device may be configured for real-time cell counting based on the whole cell count.