MSER-Based Exudate Detection in Fundus Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for detecting exudates in retinal images are inefficient and prone to human error, requiring tedious manual annotation and being sensitive to inter-image and intra-image variations, making automated detection challenging.
Innovation Solution
A novel architecture and method using maximally stable extremal regions (MSERs) for automated detection of exudates in ocular fundus images, which includes a communication interface and processing circuitry to label image regions, find extremal regions, and determine MSER ellipses parameters, allowing for robust detection without classifier training.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual annotation is used for exudate detection, then detection accuracy can be achieved, but the process is tedious and time-consuming
Solution Approach 1:
The patent replaces manual mechanical annotation with an automated computer-based system that uses image processing algorithms to detect exudates. The system automatically identifies exudate regions in retinal images without requiring manual labeling, thereby eliminating the time-consuming nature of manual annotation while maintaining detection accuracy through algorithmic analysis of image features.
Solution Approach 2:
The system enables self-service detection by allowing the computer algorithm to automatically identify and classify exudates in retinal images without human intervention. The automated processing pipeline independently performs detection, segmentation, and classification tasks that previously required manual annotation, significantly reducing processing time while maintaining diagnostic accuracy.
2Reliability
If pre-processing is applied to eliminate variations, then detection robustness improves, but processing complexity increases
Solution Approach 1:
The patent applies pre-processing operations such as contrast enhancement and normalization to retinal images before exudate detection. These preliminary actions prepare the images by reducing inter-image and intra-image variations, making the subsequent detection process more robust. The pre-processing step standardizes image characteristics, ensuring consistent detection performance across different imaging conditions.
Solution Approach 2:
The system modifies image parameters through pre-processing operations, including contrast enhancement and intensity normalization. These parameter changes adjust the image characteristics to minimize variations caused by different imaging conditions, thereby improving detection robustness. The parameter transformations make the detection algorithm more insensitive to variations in lighting, exposure, and image quality.
3Productivity
If automated processing is implemented, then productivity increases, but sensitivity to image variations becomes a problem
Solution Approach 1:
The patent implements pre-processing operations that prepare images by reducing variations before automated detection. This preliminary action ensures that the automated processing system receives standardized input, making it less sensitive to inter-image and intra-image variations. The pre-processing step includes contrast enhancement and normalization that compensate for differences in imaging conditions.
Solution Approach 2:
The system applies parameter transformations to image data, including intensity normalization and contrast adjustment, to reduce the impact of variations. These parameter changes make the automated detection algorithm more robust by transforming image characteristics into a standardized form that is less sensitive to variations in lighting, exposure, and image quality, thereby maintaining reliability while preserving high processing efficiency.
Data Source
AI summary
Architecture and a method for maximally stable extremal regions (MSERs)-based detection of exudates in an ocular fundus is disclosed. The architecture includes a communication interface to receive pixels of an ocular fundus image. The architecture further includes processing circuitry that is coupled to the communication interface. The processing circuitry is configured to automatically provide labels for light image regions and dark image regions within the ocular fundus image for a given intensity threshold and find MSERs within the ocular fundus image based on the labels. The architecture also determines MSER regions based on the MSER criteria and then highlights the pixels of the ocular fundus image that are located within MSER regions to indicate the exudates in the ocular fundus. The architecture is further configured to determine MSER ellipses parameters based on MSER regions and MSER criteria and then highlight the locations of the exudates in the ocular fundus.