Automated Edge Detection in Medical Images Using 3D Intensity Graphs

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current image analysis technologies in medicine, particularly in cardiology and echocardiography, lack fully automated systems for accurate and precise analysis of images, relying heavily on user input and manual methods, which are time-consuming and prone to observer bias, and fail to provide quantitative and qualitative diagnostic data for effective treatment planning.

Innovation Solution

A method and system for edge detection in digitized images that involves building a 3D graph of intensity distribution, calculating identifying vector angle values, setting multiple pairs of threshold level values with incremental differences, and establishing edge lines by averaging independently obtained results, allowing for automated and accurate measurement of image features like artery caliber and cardiac chamber dimensions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Extent of automation

If current edge-detection systems use first and second differentiation methods, then edge detection can be performed, but the system requires noise subtraction and frequent user input, reducing automation and increasing complexity

Engineering Contradiction:
Improveautomation of image analysisVSAvoidcomplexity of processing steps
Core Design Contradiction:
Extent of automationVSDevice complexity

Solution Approach 1:

The patent changes the fundamental parameter used for edge detection from derivative-based methods to intensity-based thresholding. By comparing pixel intensities directly against threshold values, the system eliminates the need for complex differentiation operations and noise subtraction steps, achieving full automation without increasing processing complexity

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent extracts and removes the problematic steps of noise subtraction and user-focused input from the edge detection process. By using intensity threshold comparison, the system isolates the essential edge detection function from auxiliary processing steps that require user intervention and complex noise management

Inventive Principle:
Principle #2Taking out (Extraction)

2Loss of information

If current edge-detection systems use binary edge presence/absence detection, then simple edge identification is possible, but qualitative information about different types of edges is lost

Engineering Contradiction:
Improvequalitative edge informationVSAvoidprecision of edge characterization
Core Design Contradiction:
Loss of informationVSMeasurement precision

Solution Approach 1:

The patent applies local quality by assigning different threshold values for different edge characteristics. Instead of a single binary detection, the system uses multiple intensity thresholds to characterize different types of edges (e.g., strong edges, weak edges, different orientations), preserving qualitative information about edge nature while maintaining precise measurement capability

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent transitions from binary (0 or 1) edge detection to a multi-dimensional classification system by introducing intensity threshold levels. This adds a new dimension of measurement - edge intensity strength - allowing the system to distinguish between different types of edges while maintaining precise quantitative characterization

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

3Reliability

If manual image analysis is performed by trained experts, then accurate diagnostic interpretation is possible, but observer bias and lack of reproducibility occur

Engineering Contradiction:
Improvereproducibility of diagnosisVSAvoidtime for analysis
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent implements self-service by creating an automated system that performs image analysis independently without human intervention. The intensity threshold comparison method allows the system to automatically detect and characterize edges, eliminating observer bias and ensuring reproducible results while significantly reducing the time required for diagnostic image analysis

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent replaces the mechanical system of manual expert analysis with an automated computational system. By substituting human visual inspection and interpretation with algorithmic intensity threshold comparison, the system eliminates variability in human performance while maintaining high accuracy and providing consistent, reproducible diagnostic results

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

4Measurement precision

If multiple threshold level values are used for edge detection, then more accurate edge identification is achieved, but calculation complexity increases

Engineering Contradiction:
Improveaccuracy of edge detectionVSAvoidcomplexity of threshold calculation
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent simplifies the threshold calculation by changing from complex multi-parameter optimization to a straightforward intensity-based threshold comparison. By using predetermined intensity thresholds that compare pixel values directly, the system achieves accurate edge detection while minimizing computational complexity, making the process efficient and易于 implementation

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP3555851B1Edge detection in digitized images
Publication Date: 2021.09.22 EYES LTD
  • EP3555851B1 patent drawingFigure 1
  • EP3555851B1 patent drawingFigure 2a~2b
  • EP3555851B1 patent drawingFigure 2c~2d

AI summary

A method of detecting an edge within digitized images, said method comprising steps of: (a) obtaining said image in a digital form; (b) building a 3D graph of intensity distribution within said image; (c) finding Regions of Interest ROI in the image and subtracting out the background; (d) establishing edge lines within said image; establishing a resultant edge line for each ROI by averaging or maximizing the set of independently obtained edge lines corresponding to said plurality of criteria but that differ from each other by a margin increasing with a predetermined increment; and (f) generating at least one report presenting the quantitative and/or qualitative parameters in a form of accurate quantitative measurements and qualitative inferences regarding medical and general diagnosis of the images and patients utilizing one or more output modules.