Microaneurysm Detection via Singular Spectrum Analysis Intensity Profiles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current automated systems for detecting microaneurysms in retinal images face challenges in accurately identifying these small, circular reddish dots due to low local contrast, poorly defined edges, and similarity in intensity and morphology with other retinal features, especially when located near blood vessels or in diverse ethnic backgrounds.
Innovation Solution
An image analysis method employing singular spectrum analysis (SSA) to extract intensity profiles from candidate objects, combined with multilayered dark object filtering and correlation coefficient-based scaling, to enhance feature extraction and classification, effectively distinguishing microaneurysms from other retinal structures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional template-matching or intensity profile methods are used to detect microaneurysms, then the detection process is simple, but the accuracy is insufficient due to low local contrast and similarity with other retinal features
Solution Approach 1:
The patent segments the detection process into multiple stages: candidate object identification, SSA-based intensity profile extraction, feature extraction, and classification. This multi-stage segmentation allows each stage to focus on specific aspects of the problem, improving overall detection accuracy while managing complexity through modular processing
Solution Approach 2:
The patent transitions from simple template matching in the spatial domain to singular spectrum analysis in a transformed domain. By analyzing intensity profiles across multiple dimensions and decomposing them into eigencomponents, the method extracts more discriminative features that capture the subtle characteristics of microaneurysms beyond what traditional 2D templates can detect
2Productivity
If automated detection systems are implemented to reduce workload, then screening efficiency improves, but false positives increase due to similarity between microaneurysms and other retinal structures
Solution Approach 1:
The patent implements a feedback mechanism through the classification stage, where extracted features are evaluated against learned patterns to distinguish true microaneurysms from false candidates. The system uses the extracted intensity profile features to feedback and refine the detection decisions, reducing false positives while maintaining high sensitivity
Solution Approach 2:
The patent changes the parameter space by extracting multiple features from the intensity profiles (mean, standard deviation, skewness, kurtosis, eigenvalues) rather than relying on single-parameter template matching. This multi-parameter approach provides more discriminative power to distinguish microaneurysms from similar retinal structures, improving reliability
3Measurement precision
If simple feature extraction is used, then processing speed is fast, but sensitivity to subtle microaneurysms is insufficient
Solution Approach 1:
The patent performs preliminary action by first identifying candidate objects based on basic intensity criteria before applying the more computationally intensive SSA-based feature extraction. This preliminary filtering reduces the number of candidates requiring detailed analysis, improving sensitivity for subtle microaneurysms while minimizing overall processing time through selective detailed examination
Data Source
AI summary
An image analysis method and an image processing apparatus are disclosed. The image analysis method comprises receiving a retinal image and identifying at least one candidate object from a plurality of objects within the retinal image. Singular spectrum analysis SSA is performed on the at least one candidate object, to obtain an intensity profile along at least one cross-sectional line through the candidate object. At least one feature from the intensity profile is extracted for classification of the at least one candidate object.