Pathology Slide ROI Detection from Hand-Drawn Annotations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for creating digitally annotated whole slide images in pathology are time-consuming and expensive, and hand-drawn pen annotations on glass slides are often considered artifacts and excluded from computational modeling, despite their potential value.
Innovation Solution
A computer-implemented method and tool to detect and segment hand-drawn pen annotations on digitized pathology slides, converting them into a digital format for computational models, using a pipeline that includes color space conversion, thresholding, and morphological operations to extract and refine the annotation masks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual digital annotation is performed on pathology slides, then annotation accuracy and usability for computational models is improved, but time consumption and cost increase significantly
Solution Approach 1:
The system creates a digital copy of the hand-drawn pen annotation by detecting the pen marks on the scanned slide image and converting them into digital annotation data structures. This copying process automatically transforms the physical annotation into a usable digital format without requiring manual re-annotation, thereby reducing time consumption while maintaining accuracy.
Solution Approach 2:
The system replaces the manual mechanical process of digital annotation with an automated computational process. By using image processing techniques to detect pen marks and algorithmically extract annotation regions, the system eliminates the need for manual digital drawing, significantly reducing time and cost while preserving annotation quality.
2Adaptability or versatility
If hand-drawn pen annotations are included in computational modeling, then utilization of existing annotations is improved, but annotation artifacts and noise increase
Solution Approach 1:
The system extracts only the relevant annotation information from the scanned slide by detecting pen mark locations and extracting the enclosed regions of interest. This extraction process separates the useful annotation data from the harmful artifacts (pen marks, background noise), allowing computational models to utilize the annotation intent while filtering out distracting elements.
Solution Approach 2:
The system utilizes color space conversion to differentiate between the pen annotation marks and the tissue samples. By transforming the image into different color spaces and applying color thresholding, the system isolates the pen annotation regions from the biological samples, enabling clean extraction of annotation boundaries without incorporating artifact noise.
3Measurement precision
If color space conversion is applied to differentiate annotation from ROI, then annotation detection accuracy is improved, but computational complexity increases
Solution Approach 1:
The system changes the parameter representation of pixel colors by converting between different color spaces (e.g., RGB to HSV). This parameter transformation enables better differentiation between annotation colors and tissue colors, improving detection accuracy. The computational overhead of color space conversion is manageable and provides significant accuracy benefits.
Data Source
AI summary
The present disclosure is directed to systems and methods for identifying regions of interest (ROIs) in images. A computing system may identify an image including an annotation defining an ROI. The image may have a plurality of pixels in a first color space. The computing system may convert the plurality of pixels from the first color space to a second color space to differentiate the annotation from the ROI. The computing system may select a first subset of pixels corresponding to the annotation based at least on a color value of the first subset of pixels in the second color space. The computing system may identify a second subset of pixels included in the ROI from the image using the first subset of pixels. The computing system may store an association between the second subset of pixels and the ROI defined by the annotation in the image.