Tissue Area Detection in Digital Pathology Imaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current digital pathology methods face challenges in accurately detecting tissue areas of interest from glass slides due to uneven illumination, varying staining intensity, and the presence of artifacts, which hampers efficient and accurate automated detection.
Innovation Solution
A method utilizing image segmentation techniques, including global and local statistics, contrast enhancement, and specific filtering processes to classify and distinguish tissue areas from non-tissue areas, regardless of staining methods and illumination conditions, employing two-pass segmentation and artifact removal algorithms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If automated detection is performed on thumbnail images with uneven illumination, then detection speed is improved, but detection accuracy deteriorates due to difficulty in distinguishing tissue from background
Solution Approach 1:
The patent applies local quality by using a moving window approach that analyzes local image statistics within small regions rather than treating the entire image uniformly. Each window computes local mean and standard deviation to adapt to local illumination variations, allowing the algorithm to distinguish tissue from background even when global illumination is uneven. This local analysis enables accurate detection while maintaining computational efficiency for fast processing.
2Device complexity
If threshold-based segmentation is used for AOI detection, then processing simplicity is improved, but reliability deteriorates when staining intensity varies or artifacts are present
Solution Approach 1:
The patent employs parameter changes by computing both the mean and standard deviation of pixel intensities within each moving window, rather than relying on a single threshold value. The algorithm dynamically adjusts detection criteria based on local statistical parameters (mean and standard deviation), allowing it to adapt to varying staining intensities and differentiate tissue from artifacts. This multi-parameter approach maintains computational simplicity while significantly improving detection reliability.
3Measurement precision
If comprehensive filtering is applied to remove artifacts, then detection accuracy is improved, but processing time increases
Solution Approach 1:
The patent applies segmentation by dividing the image processing into distinct stages: first computing local statistics in moving windows to identify potential tissue regions, then applying artifact filtering only to these candidate regions. This segmented approach avoids the computational burden of filtering the entire image, maintaining high detection accuracy while minimizing processing time. The segmentation allows selective application of computational resources where most needed.
Data Source
AI summary
A tissue area detection method in digital pathology imaging uses an automated method to detect tissue area of interest (AOI detection) on a Whole Slide Analysis (WSA) or Tissue Micro Array (TMA) thumbnail image. The present method may use preprocessing of the image followed by a two-pass segmentation technique for separating tissue areas from non-tissue areas. The present method may further use global and local window statistics for thresholding to overcome variations in staining intensity that may hamper accurate selection of the area of interest. The present method may also have a classifying process of the tissue area based on the staining method used and applying stain specific filters to remove unwanted artifacts.