Thyroid Nodule Boundary Clarity Detection via Multi-Index Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting the clarity of thyroid nodule boundaries in ultrasonic images are inaccurate due to limited resolution, speckle noise, and reliance on single imaging measurement indices without considering doctor judgment experience.
Innovation Solution
A method involving the acquisition of ultrasonic images, calculation of aspect ratio and ring difference coefficients, segmentation of images, and input into a Thy-Enet deep neural network and multi-layer perceptron model to determine the clarity of thyroid nodule boundaries.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single imaging measurement index is used to detect nodule boundary clarity, then the detection method is simple, but the measurement precision is insufficient
Solution Approach 1:
The patent combines multiple imaging measurement indexes (aspect ratio coefficient, inner and outer ring difference coefficient, four-partitioning intensity difference coefficient) with deep learning model outputs to form a comprehensive clarity detection system. This merging of multiple measurement approaches resolves the contradiction by achieving high measurement precision through multi-index integration while maintaining reasonable system complexity through systematic design.
Solution Approach 2:
The patent creates a composite measurement framework that integrates traditional image processing metrics (aspect ratio, intensity differences) with modern deep learning predictions. This composite approach combines the strengths of different measurement methodologies, achieving superior clarity detection accuracy by synthesizing multiple complementary measurement perspectives rather than relying on a single index.
2Loss of time
If existing clarity detection methods are used, then the detection process is quick, but the reliability of the clarity result is low
Solution Approach 1:
The patent performs preliminary calculations of multiple coefficients (aspect ratio, inner and outer ring difference, four-partitioning intensity difference) before final clarity determination. These preliminary measurements prepare comprehensive data that enhances the reliability of the final clarity assessment, allowing the system to make more accurate judgments while maintaining efficient processing through pre-computed features.
Solution Approach 2:
The patent implements a feedback mechanism where multiple measurement indexes and deep learning predictions are integrated to continuously refine the clarity determination. The system uses feedback from various measurement dimensions (geometric aspects, intensity variations, partitioned region analysis) to improve the reliability of clarity results, ensuring that the final determination is based on comprehensive evidence rather than single-metric assumptions.
3Adaptability or versatility
If doctor experience is relied upon for clarity judgment, then the detection method is flexible, but the measurement precision varies and cannot form specific standards
Solution Approach 1:
The patent enables the system to automatically perform clarity detection using objective measurement indexes and deep learning models, reducing dependency on subjective doctor experience. The system serves itself by implementing standardized algorithms that consistently evaluate clarity across different cases, eliminating the need for manual judgment while maintaining adaptability through programmable measurement protocols.
Solution Approach 2:
The patent transforms the subjective clarity assessment into objective parameter-based measurements by calculating specific coefficients (aspect ratio, intensity differences) and using deep learning predictions. This parameter transformation converts flexible but inconsistent doctor judgments into precise, standardized measurements that can be consistently applied across all cases while retaining the ability to adapt to different nodule characteristics through parameter variation.
Data Source
AI summary
A method for detecting clarity of a thyroid nodule boundary, a system thereof, an electronic device and a medium are provided. The method includes: calculating an aspect ratio coefficient; calculating an inner and outer ring difference coefficient according to an intensity average of an outer ring image and an inner ring image; segmenting the outer ring image and the inner ring image into four parts; obtaining a four-partitioning intensity difference coefficient according to the intensity average of each segmented outer ring image and each segmented inner ring image; inputting a preprocessed image into a trained Thy-Enet deep neural network to obtain the probability that the thyroid nodule boundary is clear; and inputting the aspect ratio coefficient, the inner and outer ring difference coefficient, the four-partitioning intensity difference coefficient and the probability into a trained multi-layer perceptron model to obtain a determination result of the thyroid nodule boundary.