Fluorescence Image Processing Segmentation Contrast

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current fluorescence image processing techniques fail to effectively enhance images with low signal-to-noise ratio (SNR) resulting from rapid scanning and staining times, which are necessary for high-throughput digital pathology applications, leading to inadequate contrast for human operators.

Innovation Solution

A fluorescence image processing apparatus that segments images into areas of interest and background, measures and reduces background signals, and enhances the areas of interest, using noise in the background to improve contrast, while allowing for parallel processing and storage options for iterative quality improvement.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If scanning and staining times are reduced to achieve high throughput, then productivity is improved, but signal-to-noise ratio deteriorates

Engineering Contradiction:
ImprovethroughputVSAvoidsignal-to-noise ratio
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The image is segmented into background areas and areas of interest based on intensity values. This allows different processing to be applied to different regions, enabling aggressive background suppression in low-signal regions while preserving detail in high-signal regions, thereby improving perceived SNR without requiring longer acquisition times

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Different enhancement parameters are applied to different regions of the image based on local characteristics. Background areas receive strong suppression while areas of interest receive selective enhancement. This local differentiation allows the system to maintain high productivity with reduced acquisition times while recovering SNR through region-specific processing

Inventive Principle:
Principle #3Local quality

2Manufacturing precision

If background signal is suppressed to enhance contrast, then image quality is improved, but useful information may be lost

Engineering Contradiction:
Improveimage qualityVSAvoiduseful information
Core Design Contradiction:
Manufacturing precisionVSLoss of information

Solution Approach 1:

The image is divided into background regions and foreground regions based on intensity thresholds. This segmentation allows the system to selectively suppress only the background regions while leaving the foreground regions (areas of interest) untouched or selectively enhanced, thereby improving contrast without losing useful information

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Different processing operations are applied to different spatial regions based on their characteristics. Background regions undergo strong suppression while foreground regions receive minimal or selective processing. This local differentiation ensures that contrast enhancement does not result in loss of useful information from the areas of interest

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP3077987B1Fluorescence image processing apparatus and method
Publication Date: 2017.07.05 KONINKLIJKE PHILIPS NV
  • EP3077987B1 patent drawingFigure 1~2
  • EP3077987B1 patent drawingFigure 3~4
  • EP3077987B1 patent drawingFigure 5~6

AI summary

The present invention relates to a fluorescence image processing apparatus (100) for enhancing a fluorescence image. The fluorescence image processing apparatus (100) comprises an image segmentation unit (110), a background determination unit (120), a background reduction unit (130), and an image enhancement unit (140). It is porposed to first segment the background and the area of interest of a glass slide and to use this segmentation to enhance the contrast between both. By that, scanning and staining time can be reduced, while after enhancement the image data has still a comfortable contrast for the human operator.